-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathtraining-data.txt
More file actions
1085 lines (698 loc) · 222 KB
/
Copy pathtraining-data.txt
File metadata and controls
1085 lines (698 loc) · 222 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
nix (/ˈjuːnɪks/ ⓘ, YOO-niks; trademarked as UNIX) is a family of multitasking, multi-user computer operating systems that derive from the original AT&T Unix, whose development started in 1969[1] at the Bell Labs research center by Ken Thompson, Dennis Ritchie, and others.[4] Initially intended for use inside the Bell System, AT&T licensed Unix to outside parties in the late 1970s, leading to a variety of both academic and commercial Unix variants from vendors including University of California, Berkeley (BSD), Microsoft (Xenix), Sun Microsystems (SunOS/Solaris), HP/HPE (HP-UX), and IBM (AIX).
The early versions of Unix—which are retrospectively referred to as "Research Unix"—ran on computers such as the PDP-11 and VAX; Unix was commonly used on minicomputers and mainframes from the 1970s onwards.[5] It distinguished itself from its predecessors as the first portable operating system: almost the entire operating system is written in the C programming language (in 1973), which allows Unix to operate on numerous platforms.[6] Unix systems are characterized by a modular design that is sometimes called the "Unix philosophy". According to this philosophy, the operating system should provide a set of simple tools, each of which performs a limited, well-defined function.[7] A unified and inode-based filesystem and an inter-process communication mechanism known as "pipes" serve as the main means of communication,[4] and a shell scripting and command language (the Unix shell) is used to combine the tools to perform complex workflows.
Version 7 in 1979 was the final widely released Research Unix, after which AT&T sold UNIX System III, based on Version 7, commercially in 1982; to avoid confusion between the Unix variants, AT&T combined various versions developed by others and released it as UNIX System V in 1983. However as these were closed-source, the University of California, Berkeley continued developing BSD as an alternative. Other vendors that were beginning to create commercialized versions of Unix would base their version on either System V (like Silicon Graphics's IRIX) or BSD (like SunOS). Amid the "Unix wars" of standardization, AT&T alongside Sun merged System V, BSD, SunOS and Xenix, soldifying their features into one package as UNIX System V Release 4 (SVR4) in 1989, and it was commercialized by Unix System Laboratories, an AT&T spinoff.[8][9] A rival Unix by other vendors was released as OSF/1, however most commercial Unix vendors eventually changed their distributions to be based on SVR4 with BSD features added on top.
AT&T sold Unix to Novell in 1992, who later sold the UNIX trademark to a new industry consortium called The Open Group which allow the use of the mark for certified operating systems that comply with the Single UNIX Specification (SUS).[8] Since the 1990s, Unix systems have appeared on home-class computers: BSD/OS was the first to be commercialized for i386 computers and since then free Unix-like clones of existing systems have been developed, such as FreeBSD and the combination of Linux and GNU, the latter of which have since eclipsed Unix in popularity. Unix was, until 2005, the most widely used server operating system.[10] However in the present day, Unix distributions like IBM AIX, Oracle Solaris and OpenServer continue to be widely used in certain fields.[11][12]
Overview
Version 7 Unix, the Research Unix ancestor of all modern Unix systems
Unix was originally meant to be a convenient platform for programmers developing software to be run on it and on other systems, rather than for non-programmers.[13][14][15] The system grew larger as the operating system started spreading in academic circles, and as users added their own tools to the system and shared them with colleagues.[16]
At first, Unix was not designed to support multi-tasking[17] or to be portable.[6] Later, Unix gradually gained multi-tasking and multi-user capabilities in a time-sharing configuration, as well as portability. Unix systems are characterized by various concepts: the use of plain text for storing data; a hierarchical file system; treating devices and certain types of inter-process communication (IPC) as files; and the use of a large number of software tools, small programs that can be strung together through a command-line interpreter using pipes, as opposed to using a single monolithic program that includes all of the same functionality. These concepts are collectively known as the "Unix philosophy". Brian Kernighan and Rob Pike summarize this in The Unix Programming Environment as "the idea that the power of a system comes more from the relationships among programs than from the programs themselves".[18]
By the early 1980s, users began seeing Unix as a potential universal operating system, suitable for computers of all sizes.[19][20] The Unix environment and the client–server program model were essential elements in the development of the Internet and the reshaping of computing as centered in networks rather than in individual computers.
Both Unix and the C programming language were developed by AT&T and distributed to government and academic institutions, which led to both being ported to a wider variety of machine families than any other operating system.
The Unix operating system consists of many libraries and utilities along with the master control program, the kernel. The kernel provides services to start and stop programs, handles the file system and other common "low-level" tasks that most programs share, and schedules access to avoid conflicts when programs try to access the same resource or device simultaneously. To mediate such access, the kernel has special rights, reflected in the distinction of kernel space from user space, the latter being a lower priority realm where most application programs operate.
History
Main article: History of Unix
The origins of Unix date back to the mid-1960s when the Massachusetts Institute of Technology, Bell Labs, and General Electric were developing Multics, a time-sharing operating system for the GE 645 mainframe computer.[21] Multics featured several innovations, but also presented severe problems. Frustrated by the size and complexity of Multics, but not by its goals, individual researchers at Bell Labs started withdrawing from the project. The last to leave were Ken Thompson, Dennis Ritchie, Douglas McIlroy, and Joe Ossanna,[17] who decided to reimplement their experiences in a new project of smaller scale. This new operating system was initially without organizational backing, and also without a name.
The new operating system was a single-tasking system.[17] In 1970, the group coined the name Unics for Uniplexed Information and Computing Service as a pun on Multics, which stood for Multiplexed Information and Computer Services. Brian Kernighan takes credit for the idea, but adds that "no one can remember" the origin of the final spelling Unix.[22] Dennis Ritchie,[17] Doug McIlroy,[1] and Peter G. Neumann[23] also credit Kernighan.
The operating system was originally written in assembly language, but in 1973, Version 4 Unix was rewritten in C. Ken Thompson faced multiple challenges attempting the kernel port due to the evolving state of C, which lacked key features like structures at the time.[17][24] Version 4 Unix, however, still had much PDP-11 specific code, and was not suitable for porting. The first port to another platform was a port of Version 6, made four years later (1977) at the University of Wollongong for the Interdata 7/32,[25] followed by a Bell Labs port of Version 7 to the Interdata 8/32 during 1977 and 1978.[26]
Bell Labs produced several versions of Unix that are collectively referred to as Research Unix. In 1975, the first source license for UNIX was sold to Donald B. Gillies at the University of Illinois Urbana–Champaign (UIUC) Department of Computer Science.[27]
During the late 1970s and early 1980s, the influence of Unix in academic circles led to large-scale adoption of Unix (BSD and System V) by commercial startups, which in turn led to Unix fragmenting into multiple, similar — but often slightly and mutually incompatible — systems including DYNIX, HP-UX, SunOS/Solaris, AIX, and Xenix. In the late 1980s, AT&T Unix System Laboratories and Sun Microsystems developed System V Release 4 (SVR4), which was subsequently adopted by many commercial Unix vendors.
In the 1990s, Unix and Unix-like systems grew in popularity and became the operating system of choice for over 90% of the world's top 500 fastest supercomputers,[28] as BSD and Linux distributions were developed through collaboration by a worldwide network of programmers. In 2000, Apple released Darwin, also a Unix system, which became the core of the Mac OS X operating system, later renamed macOS.[29]
Unix-like operating systems are widely used in modern servers, workstations, and mobile devices.[30]
Standards
The Common Desktop Environment (CDE), part of the COSE initiative
In the late 1980s, an open operating system standardization effort now known as POSIX provided a common baseline for all operating systems; IEEE based POSIX around the common structure of the major competing variants of the Unix system, publishing the first POSIX standard in 1988. In the early 1990s, a separate but very similar effort was started by an industry consortium, the Common Open Software Environment (COSE) initiative, which eventually became the Single UNIX Specification (SUS) administered by The Open Group. Starting in 1998, the Open Group and IEEE started the Austin Group, to provide a common definition of POSIX and the Single UNIX Specification, which, by 2008, had become the Open Group Base Specification.
In 1999, in an effort towards compatibility, several Unix system vendors agreed on SVR4's Executable and Linkable Format (ELF) as the standard for binary and object code files. The common format allows substantial binary compatibility among different Unix systems operating on the same CPU architecture.
The Filesystem Hierarchy Standard was created to provide a reference directory layout for Unix-like operating systems; it has mainly been used in Linux.
Components
See also: List of Unix commands
This section needs additional citations for verification. Please help improve this article by adding citations to reliable sources in this section. Unsourced material may be challenged and removed.
Find sources: "Unix" – news · newspapers · books · scholar · JSTOR (October 2023) (Learn how and when to remove this message)
The Unix system is composed of several components that were originally packaged together. By including the development environment, libraries, documents and the portable, modifiable source code for all of these components, in addition to the kernel of an operating system, Unix was a self-contained software system. This was one of the key reasons it emerged as an important teaching and learning tool and has had a broad influence. See § Impact, below.
The inclusion of these components did not make the system large – the original V7 UNIX distribution, consisting of copies of all of the compiled binaries plus all of the source code and documentation occupied less than 10 MB and arrived on a single nine-track magnetic tape, earning its reputation as a portable system.[31] The printed documentation, typeset from the online sources, was contained in two volumes.
The names and filesystem locations of the Unix components have changed substantially across the history of the system. Nonetheless, the V7 implementation has the canonical early structure:
Kernel – source code in /usr/sys, composed of several sub-components:
conf – configuration and machine-dependent parts, including boot code
dev – device drivers for control of hardware (and some pseudo-hardware)
sys – operating system "kernel", handling memory management, process scheduling, system calls, etc.
h – header files, defining key structures within the system and important system-specific invariables
Development environment – early versions of Unix contained a development environment sufficient to recreate the entire system from source code:
ed – text editor, for creating source code files
cc – C language compiler (first appeared in V3 Unix)
as – machine-language assembler for the machine
ld – linker, for combining object files
lib – object-code libraries (installed in /lib or /usr/lib). libc, the system library with C run-time support, was the primary library, but there have always been additional libraries for things such as mathematical functions (libm) or database access. V7 Unix introduced the first version of the modern "Standard I/O" library stdio as part of the system library. Later implementations increased the number of libraries significantly.
make – build manager (introduced in PWB/UNIX), for effectively automating the build process
include – header files for software development, defining standard interfaces and system invariants
Other languages – V7 Unix contained a Fortran-77 compiler, a programmable arbitrary-precision calculator (bc, dc), and the awk scripting language; later versions and implementations contain many other language compilers and toolsets. Early BSD releases included Pascal tools, and many modern Unix systems also include the GNU Compiler Collection as well as or instead of a proprietary compiler system.
Other tools – including an object-code archive manager (ar), symbol-table lister (nm), compiler-development tools (e.g. lex & yacc), and debugging tools.
Commands – Unix makes little distinction between commands (user-level programs) for system operation and maintenance (e.g. cron), commands of general utility (e.g. grep), and more general-purpose applications such as the text formatting and typesetting package. Nonetheless, some major categories are:
sh – the "shell" programmable command-line interpreter, the primary user interface on Unix before window systems appeared, and even afterward (within a "command window").
Utilities – the core toolkit of the Unix command set, including cp, ls, grep, find and many others. Subcategories include:
System utilities – administrative tools such as mkfs, fsck, and many others.
User utilities – environment management tools such as passwd, kill, and others.
Document formatting – Unix systems were used from the outset for document preparation and typesetting systems, and included many related programs such as nroff, troff, tbl, eqn, refer, and pic. Some modern Unix systems also include packages such as TeX and Ghostscript.
Graphics – the plot subsystem provided facilities for producing simple vector plots in a device-independent format, with device-specific interpreters to display such files. Modern Unix systems also generally include X11 as a standard windowing system and GUI, and many support OpenGL.
Communications – early Unix systems contained no inter-system communication, but did include the inter-user communication programs mail and write. V7 introduced the early inter-system communication system UUCP, and systems beginning with BSD release 4.1c included TCP/IP utilities.
Documentation – Unix was one of the first operating systems to include all of its documentation online in machine-readable form.[32] The documentation included:
man – manual pages for each command, library component, system call, header file, etc.
doc – longer documents detailing major subsystems, such as the C language and troff
Impact
See also: Unix-like
Ken Thompson and Dennis Ritchie, principal developers of Research Unix
Photo from USENIX 1984, including Dennis Ritchie (center)
The Unix system had a significant impact on other operating systems. It achieved its reputation by its interactivity, by providing the software at a nominal fee for educational use, by running on inexpensive hardware, and by being easy to adapt and move to different machines. Unix was originally written in assembly language, but was soon rewritten in C, a high-level programming language.[33] Although this followed the lead of CTSS, Multics and Burroughs MCP, it was Unix that popularized the idea.
Unix had a drastically simplified file model compared to many contemporary operating systems: treating all kinds of files as simple byte arrays. The file system hierarchy contained machine services and devices (such as printers, terminals, or disk drives), providing a uniform interface, but at the expense of occasionally requiring additional mechanisms such as ioctl and mode flags to access features of the hardware that did not fit the simple "stream of bytes" model. The Plan 9 operating system pushed this model even further and eliminated the need for additional mechanisms.
Unix also popularized the hierarchical file system with arbitrarily nested subdirectories, originally introduced by Multics. Other common operating systems of the era had ways to divide a storage device into multiple directories or sections, but they had a fixed number of levels, often only one level. Several major proprietary operating systems eventually added recursive subdirectory capabilities also patterned after Multics. DEC's RSX-11M's "group, user" hierarchy evolved into OpenVMS directories, CP/M's volumes evolved into MS-DOS 2.0+ subdirectories, and HP's MPE group.account hierarchy and IBM's SSP and OS/400 library systems were folded into broader POSIX file systems.
Making the command interpreter an ordinary user-level program, with additional commands provided as separate programs, was another Multics innovation popularized by Unix. The Unix shell used the same language for interactive commands as for scripting (shell scripts – there was no separate job control language like IBM's JCL). Since the shell and OS commands were "just another program", the user could choose (or even write) their own shell. New commands could be added without changing the shell itself. Unix's innovative command-line syntax for creating modular chains of producer-consumer processes (pipelines) made a powerful programming paradigm (coroutines) widely available. Many later command-line interpreters have been inspired by the Unix shell.
A fundamental simplifying assumption of Unix was its focus on newline-delimited text for nearly all file formats. There were no "binary" editors in the original version of Unix – the entire system was configured using textual shell command scripts. The common denominator in the I/O system was the byte – unlike "record-based" file systems. The focus on text for representing nearly everything made Unix pipes especially useful and encouraged the development of simple, general tools that could easily be combined to perform more complicated ad hoc tasks. The focus on text and bytes made the system far more scalable and portable than other systems. Over time, text-based applications have also proven popular in application areas, such as printing languages (PostScript, ODF), and at the application layer of the Internet protocols, e.g., FTP, SMTP, HTTP, SOAP, and SIP.
Unix popularized a syntax for regular expressions that found widespread use. The Unix programming interface became the basis for a widely implemented operating system interface standard (POSIX, see above). The C programming language soon spread beyond Unix, and is now ubiquitous in systems and applications programming.
Early Unix developers were important in bringing the concepts of modularity and reusability into software engineering practice, spawning a "software tools" movement. Over time, the leading developers of Unix (and programs that ran on it) established a set of cultural norms for developing software, norms which became as important and influential as the technology of Unix itself; this has been termed the Unix philosophy.
The TCP/IP networking protocols were quickly implemented on the Unix versions widely used on relatively inexpensive computers, which contributed to the Internet explosion of worldwide, real-time connectivity and formed the basis for implementations on many other platforms.
The Unix policy of extensive on-line documentation and (for many years) ready access to all system source code raised programmer expectations, and contributed to the launch of the free software movement in 1983.
Free Unix and Unix-like variants
See also: Operating system § Unix and Unix-like operating systems
Console screenshots of Debian (top, a popular Linux distribution) and FreeBSD (bottom, a popular Unix-like operating system)
In 1983, Richard Stallman announced the GNU (short for "GNU's Not Unix") project, an ambitious effort to create a free software Unix-like system—"free" in the sense that everyone who received a copy would be free to use, study, modify, and redistribute it. The GNU project's own kernel development project, GNU Hurd, had not yet produced a working kernel, but in 1991 Linus Torvalds released the Linux kernel as free software under the GNU General Public License. In addition to their use in the GNU operating system, many GNU packages – such as the GNU Compiler Collection (and the rest of the GNU toolchain), the GNU C library and the GNU Core Utilities – have gone on to play central roles in other free Unix systems as well.
Linux distributions, consisting of the Linux kernel and large collections of compatible software have become popular both with individual users and in business. Popular distributions include Red Hat Enterprise Linux, Fedora, SUSE Linux Enterprise, openSUSE, Debian, Ubuntu, Linux Mint, Slackware Linux, Arch Linux and Gentoo.[34]
A free derivative of BSD Unix, 386BSD, was released in 1992 and led to the NetBSD and FreeBSD projects. With the 1994 settlement of a lawsuit brought against the University of California and Berkeley Software Design Inc. (USL v. BSDi) by Unix System Laboratories, it was clarified that Berkeley had the right to distribute BSD Unix for free if it so desired. Since then, BSD Unix has been developed in several different product branches, including OpenBSD and DragonFly BSD.
Because of the modular design of the Unix model, sharing components is relatively common: most or all Unix and Unix-like systems include at least some BSD code, while some include GNU utilities in their distributions. Linux and BSD Unix are increasingly filling market needs traditionally served by proprietary Unix operating systems, expanding into new markets such as the consumer desktop, mobile devices and embedded devices.
In a 1999 interview, Dennis Ritchie voiced his opinion that Linux and BSD Unix operating systems are a continuation of the basis of the Unix design and are derivatives of Unix:[35]
I think the Linux phenomenon is quite delightful, because it draws so strongly on the basis that Unix provided. Linux seems to be among the healthiest of the direct Unix derivatives, though there are also the various BSD systems as well as the more official offerings from the workstation and mainframe manufacturers.
In the same interview, he states that he views both Unix and Linux as "the continuation of ideas that were started by Ken and me and many others, many years ago".[35]
OpenSolaris was the free software counterpart to Solaris developed by Sun Microsystems, which included a CDDL-licensed kernel and a primarily GNU userland. However, Oracle discontinued the project upon their acquisition of Sun, which prompted a group of former Sun employees and members of the OpenSolaris community to fork OpenSolaris into the illumos kernel. As of 2014, illumos remains the only active, open-source System V derivative.
ARPANET
In May 1975, RFC 681 described the development of Network Unix by the Center for Advanced Computation at the University of Illinois Urbana-Champaign.[36] The Unix system was said to "present several interesting capabilities as an ARPANET mini-host". At the time, Unix required a license from Bell Telephone Laboratories that cost US$20,000 for non-university institutions, while universities could obtain a license for a nominal fee of $150. It was noted that Bell was "open to suggestions" for an ARPANET-wide license.
The RFC specifically mentions that Unix "offers powerful local processing facilities in terms of user programs, several compilers, an editor based on QED, a versatile document preparation system, and an efficient file system featuring sophisticated access control, mountable and de-mountable volumes, and a unified treatment of peripherals as special files." The latter permitted the Network Control Program (NCP) to be integrated within the Unix file system, treating network connections as special files that could be accessed through standard Unix I/O calls, which included the added benefit of closing all connections on program exit, should the user neglect to do so. In order "to minimize the amount of code added to the basic Unix kernel", much of the NCP code ran in a swappable user process, running only when needed.[36]
Branding
See also: List of Unix systems
Promotional license plate by Digital Equipment Corporation. Actual license plate is used by Jon Hall.
HP 9000 workstation running HP-UX, a certified Unix operating system
AT&T originally did not allow licensees to use the Unix name; thus Microsoft called its variant Xenix, for example.[37] In October 1988, they allowed licensees to use the UNIX trademark for systems based on System V Release 3.2, if certain conditions were met.[38] In October 1993, Novell, the company that owned the rights to the Unix System V source at the time, transferred the trademarks of Unix to the X/Open Company (now The Open Group),[39] and in 1995 sold the related business operations to Santa Cruz Operation (SCO).[40][41] Whether Novell also sold the copyrights to the actual software was the subject of a federal lawsuit in 2006, SCO v. Novell, which Novell won. The case was appealed, but on August 30, 2011, the United States Court of Appeals for the Tenth Circuit affirmed the trial decisions, closing the case.[42] Unix vendor SCO Group Inc. accused Novell of slander of title.
The present owner of the trademark UNIX is The Open Group, an industry standards consortium. Only systems fully compliant with and certified to the Single UNIX Specification qualify as "UNIX" (others are called "Unix-like").
By decree of The Open Group, the term "UNIX" refers more to a class of operating systems than to a specific implementation of an operating system; those operating systems which meet The Open Group's Single UNIX Specification should be able to bear the UNIX 98 or UNIX 03 trademarks today, after the operating system's vendor pays a substantial certification fee and annual trademark royalties to The Open Group.[43] Systems that have been licensed to use the UNIX trademark include AIX,[44] EulerOS,[45] HP-UX,[46] Inspur K-UX,[47] IRIX,[48] macOS,[49] Solaris,[50] Tru64 UNIX (formerly "Digital UNIX", or OSF/1),[51] and z/OS.[52] Notably, EulerOS and Inspur K-UX are Linux distributions certified as UNIX 03 compliant.[53][54]
Sometimes a representation like Un*x, *NIX, or *N?X is used to indicate all operating systems similar to Unix. This comes from the use of the asterisk (*) and the question mark characters as wildcard indicators in many utilities. This notation is also used to describe other Unix-like systems that have not met the requirements for UNIX branding from the Open Group.
The Open Group requests that UNIX always be used as an adjective followed by a generic term such as system to help avoid the creation of a genericized trademark.
Unix was the original formatting,[disputed – discuss] but the usage of UNIX remains widespread because it was once typeset in small caps (Unix). According to Dennis Ritchie, when presenting the original Unix paper to the third Operating Systems Symposium of the American Association for Computing Machinery (ACM), "we had a new typesetter and troff had just been invented and we were intoxicated by being able to produce small caps".[55] Many of the operating system's predecessors and contemporaries used all-uppercase lettering, so many people wrote the name in upper case due to force of habit. It is not an acronym.[56]
Trademark names can be registered by different entities in different countries and trademark laws in some countries allow the same trademark name to be controlled by two different entities if each entity uses the trademark in easily distinguishable categories. The result is that Unix has been used as a brand name for various products including bookshelves, ink pens, bottled glue, diapers, hair driers and food containers.[57]
Several plural forms of Unix are used casually to refer to multiple brands of Unix and Unix-like systems. Most common is the conventional Unixes, but Unices, treating Unix as a Latin noun of the third declension, is also popular. The pseudo-Anglo-Saxon plural form Unixen is not common, although occasionally seen. Sun Microsystems, developer of the Solaris variant, has asserted that the term Unix is itself plural, referencing its many implementations.[58]
A large language model (LLM) is a type of machine learning model designed for natural language processing tasks such as language generation. LLMs are language models with many parameters, and are trained with self-supervised learning on a vast amount of text.
The largest and most capable LLMs are generative pretrained transformers (GPTs). Modern models can be fine-tuned for specific tasks or guided by prompt engineering.[1] These models acquire predictive power regarding syntax, semantics, and ontologies[2] inherent in human language corpora, but they also inherit inaccuracies and biases present in the data they are trained in.[3]
History
The training compute of notable large models in FLOPs vs publication date over the period 2010-2024. For overall notable models (top left), frontier models (top right), top language models (bottom left) and top models within leading companies (bottom right). The majority of these models are language models.
The training compute of notable large AI models in FLOPs vs publication date over the period 2017-2024. The majority of large models are language models or multimodal models with language capacity.
Before 2017, there were a few language models that were large as compared to capacities then available. In the 1990s, the IBM alignment models pioneered statistical language modelling. A smoothed n-gram model in 2001 trained on 0.3 billion words achieved state-of-the-art perplexity at the time.[4] In the 2000s, as Internet use became prevalent, some researchers constructed Internet-scale language datasets ("web as corpus"[5]), upon which they trained statistical language models.[6][7] In 2009, in most language processing tasks, statistical language models dominated over symbolic language models because they can usefully ingest large datasets.[8]
After neural networks became dominant in image processing around 2012,[9] they were applied to language modelling as well. Google converted its translation service to Neural Machine Translation in 2016. Because it preceded the existence of transformers, it was done by seq2seq deep LSTM networks.
An illustration of main components of the transformer model from the original paper, where layers were normalized after (instead of before) multiheaded attention
At the 2017 NeurIPS conference, Google researchers introduced the transformer architecture in their landmark paper "Attention Is All You Need". This paper's goal was to improve upon 2014 seq2seq technology,[10] and was based mainly on the attention mechanism developed by Bahdanau et al. in 2014.[11] The following year in 2018, BERT was introduced and quickly became "ubiquitous".[12] Though the original transformer has both encoder and decoder blocks, BERT is an encoder-only model. Academic and research usage of BERT began to decline in 2023, following rapid improvements in the abilities of decoder-only models (such as GPT) to solve tasks via prompting.[13]
Although decoder-only GPT-1 was introduced in 2018, it was GPT-2 in 2019 that caught widespread attention because OpenAI at first deemed it too powerful to release publicly, out of fear of malicious use.[14] GPT-3 in 2020 went a step further and as of 2024 is available only via API with no offering of downloading the model to execute locally. But it was the 2022 consumer-facing browser-based ChatGPT that captured the imaginations of the general population and caused some media hype and online buzz.[15] The 2023 GPT-4 was praised for its increased accuracy and as a "holy grail" for its multimodal capabilities.[16] OpenAI did not reveal the high-level architecture and the number of parameters of GPT-4. The release of ChatGPT led to an uptick in LLM usage across several research subfields of computer science, including robotics, software engineering, and societal impact work.[17] In 2024 OpenAI released the reasoning model OpenAI o1, which generates long chains of thought before returning a final answer.
Competing language models have for the most part been attempting to equal the GPT series, at least in terms of number of parameters.[18]
Since 2022, source-available models have been gaining popularity, especially at first with BLOOM and LLaMA, though both have restrictions on the field of use. Mistral AI's models Mistral 7B and Mixtral 8x7b have the more permissive Apache License. In January 2025, DeepSeek released DeepSeek R1, a 671-billion-parameter open-weight model that performs comparably to OpenAI o1 but at a much lower cost.[19]
Since 2023, many LLMs have been trained to be multimodal, having the ability to also process or generate other types of data, such as images or audio. These LLMs are also called large multimodal models (LMMs).[20]
As of 2024, the largest and most capable models are all based on the transformer architecture. Some recent implementations are based on other architectures, such as recurrent neural network variants and Mamba (a state space model).[21][22][23]
Dataset preprocessing
See also: List of datasets for machine-learning research § Internet
Tokenization
As machine learning algorithms process numbers rather than text, the text must be converted to numbers. In the first step, a vocabulary is decided upon, then integer indices are arbitrarily but uniquely assigned to each vocabulary entry, and finally, an embedding is associated to the integer index. Algorithms include byte-pair encoding (BPE) and WordPiece. There are also special tokens serving as control characters, such as [MASK] for masked-out token (as used in BERT), and [UNK] ("unknown") for characters not appearing in the vocabulary. Also, some special symbols are used to denote special text formatting. For example, "Ġ" denotes a preceding whitespace in RoBERTa and GPT. "##" denotes continuation of a preceding word in BERT.[24]
For example, the BPE tokenizer used by GPT-3 (Legacy) would split tokenizer: texts -> series of numerical "tokens" as
token izer : texts -> series of numerical " t ok ens "
Tokenization also compresses the datasets. Because LLMs generally require input to be an array that is not jagged, the shorter texts must be "padded" until they match the length of the longest one. How many tokens are, on average, needed per word depends on the language of the dataset.[25][26]
BPE
Main article: Byte pair encoding
As an example, consider a tokenizer based on byte-pair encoding. In the first step, all unique characters (including blanks and punctuation marks) are treated as an initial set of n-grams (i.e. initial set of uni-grams). Successively the most frequent pair of adjacent characters is merged into a bi-gram and all instances of the pair are replaced by it. All occurrences of adjacent pairs of (previously merged) n-grams that most frequently occur together are then again merged into even lengthier n-gram, until a vocabulary of prescribed size is obtained (in case of GPT-3, the size is 50257).[27] After a tokenizer is trained, any text can be tokenized by it, as long as it does not contain characters not appearing in the initial-set of uni-grams.[28]
Problems
A token vocabulary based on the frequencies extracted from mainly English corpora uses as few tokens as possible for an average English word. However, an average word in another language encoded by such an English-optimized tokenizer is split into a suboptimal amount of tokens. GPT-2 tokenizer can use up to 15 times more tokens per word for some languages, for example for the Shan language from Myanmar. Even more widespread languages such as Portuguese and German have "a premium of 50%" compared to English.[26]
Greedy tokenization also causes subtle problems with text completion.[29]
Dataset cleaning
Main article: Data cleansing
In the context of training LLMs, datasets are typically cleaned by removing low-quality, duplicated, or toxic data.[30] Cleaned datasets can increase training efficiency and lead to improved downstream performance.[31][32] A trained LLM can be used to clean datasets for training a further LLM.[33]
With the increasing proportion of LLM-generated content on the web, data cleaning in the future may include filtering out such content. LLM-generated content can pose a problem if the content is similar to human text (making filtering difficult) but of lower quality (degrading performance of models trained on it).[34]
Synthetic data
Main article: Synthetic data
Training of largest language models might need more linguistic data than naturally available, or that the naturally occurring data is of insufficient quality. In these cases, synthetic data might be used. Microsoft's Phi series of LLMs is trained on textbook-like data generated by another LLM.[35]
Training and architecture
See also: Fine-tuning (machine learning)
Reinforcement learning from human feedback
Reinforcement learning from human feedback (RLHF) through algorithms, such as proximal policy optimization, is used to further fine-tune a model based on a dataset of human preferences.[36]
Instruction tuning
Using "self-instruct" approaches, LLMs have been able to bootstrap correct responses, replacing any naive responses, starting from human-generated corrections of a few cases. For example, in the instruction "Write an essay about the main themes represented in Hamlet," an initial naive completion might be "If you submit the essay after March 17, your grade will be reduced by 10% for each day of delay," based on the frequency of this textual sequence in the corpus.[37]
Mixture of experts
Main article: Mixture of experts
The largest LLM may be too expensive to train and use directly. For such models, mixture of experts (MoE) can be applied, a line of research pursued by Google researchers since 2017 to train models reaching up to 1 trillion parameters.[38][39][40]
Prompt engineering, attention mechanism, and context window
See also: Prompt engineering and Attention (machine learning)
Most results previously achievable only by (costly) fine-tuning, can be achieved through prompt engineering, although limited to the scope of a single conversation (more precisely, limited to the scope of a context window).[41]
When each head calculates, according to its own criteria, how much other tokens are relevant for the "it_" token, note that the second attention head, represented by the second column, is focusing most on the first two rows, i.e. the tokens "The" and "animal", while the third column is focusing most on the bottom two rows, i.e. on "tired", which has been tokenized into two tokens.[42]
In order to find out which tokens are relevant to each other within the scope of the context window, the attention mechanism calculates "soft" weights for each token, more precisely for its embedding, by using multiple attention heads, each with its own "relevance" for calculating its own soft weights. For example, the small (i.e. 117M parameter sized) GPT-2 model has had twelve attention heads and a context window of only 1k tokens.[43] In its medium version it has 345M parameters and contains 24 layers, each with 12 attention heads. For the training with gradient descent a batch size of 512 was utilized.[28]
The largest models, such as Google's Gemini 1.5, presented in February 2024, can have a context window sized up to 1 million (context window of 10 million was also "successfully tested").[44] Other models with large context windows includes Anthropic's Claude 2.1, with a context window of up to 200k tokens.[45] Note that this maximum refers to the number of input tokens and that the maximum number of output tokens differs from the input and is often smaller. For example, the GPT-4 Turbo model has a maximum output of 4096 tokens.[46]
Length of a conversation that the model can take into account when generating its next answer is limited by the size of a context window, as well. If the length of a conversation, for example with ChatGPT, is longer than its context window, only the parts inside the context window are taken into account when generating the next answer, or the model needs to apply some algorithm to summarize the too distant parts of conversation.
The shortcomings of making a context window larger include higher computational cost and possibly diluting the focus on local context, while making it smaller can cause a model to miss an important long-range dependency. Balancing them is a matter of experimentation and domain-specific considerations.
A model may be pre-trained either to predict how the segment continues, or what is missing in the segment, given a segment from its training dataset.[47] It can be either
autoregressive (i.e. predicting how the segment continues, as GPTs do): for example given a segment "I like to eat", the model predicts "ice cream", or "sushi".
"masked" (i.e. filling in the parts missing from the segment, the way "BERT"[48] does it): for example, given a segment "I like to [__] [__] cream", the model predicts that "eat" and "ice" are missing.
Models may be trained on auxiliary tasks which test their understanding of the data distribution, such as Next Sentence Prediction (NSP), in which pairs of sentences are presented and the model must predict whether they appear consecutively in the training corpus.[48] During training, regularization loss is also used to stabilize training. However regularization loss is usually not used during testing and evaluation.
Infrastructure
Substantial infrastructure is necessary for training the largest models.[49][50][51]
Training cost
The qualifier "large" in "large language model" is inherently vague, as there is no definitive threshold for the number of parameters required to qualify as "large". As time goes on, what was previously considered "large" may evolve. GPT-1 of 2018 is usually considered the first LLM, even though it has only 0.117 billion parameters. The tendency towards larger models is visible in the list of large language models.
As technology advanced, large sums have been invested in increasingly large models. For example, training of the GPT-2 (i.e. a 1.5-billion-parameters model) in 2019 cost $50,000, while training of the PaLM (i.e. a 540-billion-parameters model) in 2022 cost $8 million, and Megatron-Turing NLG 530B (in 2021) cost around $11 million.[52]
For Transformer-based LLM, training cost is much higher than inference cost. It costs 6 FLOPs per parameter to train on one token, whereas it costs 1 to 2 FLOPs per parameter to infer on one token.[53]
Tool use
There are certain tasks that, in principle, cannot be solved by any LLM, at least not without the use of external tools or additional software. An example of such a task is responding to the user's input '354 * 139 = ', provided that the LLM has not already encountered a continuation of this calculation in its training corpus.[dubious – discuss] In such cases, the LLM needs to resort to running program code that calculates the result, which can then be included in its response.[dubious – discuss]: Another example is "What is the time now? It is ", where a separate program interpreter would need to execute a code to get system time on the computer, so that the LLM can include it in its reply.[54][55] This basic strategy can be sophisticated with multiple attempts of generated programs, and other sampling strategies.[56]
Generally, in order to get an LLM to use tools, one must fine-tune it for tool-use. If the number of tools is finite, then fine-tuning may be done just once. If the number of tools can grow arbitrarily, as with online API services, then the LLM can be fine-tuned to be able to read API documentation and call API correctly.[57][58]
Retrieval-augmented generation (RAG) is another approach that enhances LLMs by integrating them with document retrieval systems. Given a query, a document retriever is called to retrieve the most relevant documents. This is usually done by encoding the query and the documents into vectors, then finding the documents with vectors (usually stored in a vector database) most similar to the vector of the query. The LLM then generates an output based on both the query and context included from the retrieved documents.[59]
Agency
An LLM is typically not an autonomous agent by itself, as it lacks the ability to interact with dynamic environments, recall past behaviors, and plan future actions, but can be transformed into one by integrating modules like profiling, memory, planning, and action.[60]
The ReAct pattern, a portmanteau of "Reason + Act", constructs an agent out of an LLM, using the LLM as a planner. The LLM is prompted to "think out loud". Specifically, the language model is prompted with a textual description of the environment, a goal, a list of possible actions, and a record of the actions and observations so far. It generates one or more thoughts before generating an action, which is then executed in the environment.[61] The linguistic description of the environment given to the LLM planner can even be the LaTeX code of a paper describing the environment.[62]
In the DEPS ("Describe, Explain, Plan and Select") method, an LLM is first connected to the visual world via image descriptions, then it is prompted to produce plans for complex tasks and behaviors based on its pretrained knowledge and environmental feedback it receives.[63]
The Reflexion method[64] constructs an agent that learns over multiple episodes. At the end of each episode, the LLM is given the record of the episode, and prompted to think up "lessons learned", which would help it perform better at a subsequent episode. These "lessons learned" are given to the agent in the subsequent episodes.[citation needed]
Monte Carlo tree search can use an LLM as rollout heuristic. When a programmatic world model is not available, an LLM can also be prompted with a description of the environment to act as world model.[65]
For open-ended exploration, an LLM can be used to score observations for their "interestingness", which can be used as a reward signal to guide a normal (non-LLM) reinforcement learning agent.[66] Alternatively, it can propose increasingly difficult tasks for curriculum learning.[67] Instead of outputting individual actions, an LLM planner can also construct "skills", or functions for complex action sequences. The skills can be stored and later invoked, allowing increasing levels of abstraction in planning.[67]
LLM-powered agents can keep a long-term memory of its previous contexts, and the memory can be retrieved in the same way as Retrieval Augmented Generation. Multiple such agents can interact socially.[68]
Compression
Typically, LLMs are trained with single- or half-precision floating point numbers (float32 and float16). One float16 has 16 bits, or 2 bytes, and so one billion parameters require 2 gigabytes. The largest models typically have 100 billion parameters, requiring 200 gigabytes to load, which places them outside the range of most consumer electronics.[69]
Post-training quantization[70] aims to decrease the space requirement by lowering precision of the parameters of a trained model, while preserving most of its performance.[71][72] The simplest form of quantization simply truncates all numbers to a given number of bits. It can be improved by using a different quantization codebook per layer. Further improvement can be done by applying different precisions to different parameters, with higher precision for particularly important parameters ("outlier weights").[73] See the visual guide to quantization by Maarten Grootendorst[74] for a visual depiction.
While quantized models are typically frozen, and only pre-quantized models are fine-tuned, quantized models can still be fine-tuned.[75]
Multimodality
See also: Multimodal learning
Multimodality means "having several modalities", and a "modality" refers to a type of input or output, such as video, image, audio, text, proprioception, etc.[76] There have been many AI models trained specifically to ingest one modality and output another modality, such as AlexNet for image to label,[77] visual question answering for image-text to text,[78] and speech recognition for speech to text.
A common method to create multimodal models out of an LLM is to "tokenize" the output of a trained encoder. Concretely, one can construct an LLM that can understand images as follows: take a trained LLM, and take a trained image encoder E {\displaystyle E}. Make a small multilayered perceptron f {\displaystyle f}, so that for any image y {\displaystyle y}, the post-processed vector f ( E ( y ) ) {\displaystyle f(E(y))} has the same dimensions as an encoded token. That is an "image token". Then, one can interleave text tokens and image tokens. The compound model is then fine-tuned on an image-text dataset. This basic construction can be applied with more sophistication to improve the model. The image encoder may be frozen to improve stability.[79]
Flamingo demonstrated the effectiveness of the tokenization method, finetuning a pair of pretrained language model and image encoder to perform better on visual question answering than models trained from scratch.[80] Google PaLM model was fine-tuned into a multimodal model PaLM-E using the tokenization method, and applied to robotic control.[81] LLaMA models have also been turned multimodal using the tokenization method, to allow image inputs,[82] and video inputs.[83]
GPT-4 can use both text and image as inputs[84] (although the vision component was not released to the public until GPT-4V[85]); Google DeepMind's Gemini is also multimodal.[86] Mistral introduced its own multimodel Pixtral 12B model in September 2024.[87]
Reasoning
In late 2024, a new direction emerged in LLM development with models specifically designed for complex reasoning tasks. These "reasoning models" were trained to spend more time generating step-by-step solutions before providing final answers, similar to human problem-solving processes.[88] OpenAI introduced this trend with their o1 model in September 2024, followed by o3 in December 2024. These models showed significant improvements in mathematics, science, and coding tasks compared to traditional LLMs. For example, on International Mathematics Olympiad qualifying exam problems, GPT-4o achieved 13% accuracy while o1 reached 83%.[88][89] In January 2025, the Chinese company DeepSeek released DeepSeek-R1, a 671-billion-parameter open-weight reasoning model that achieved comparable performance to OpenAI's o1 while being significantly more cost-effective to operate. Unlike proprietary models from OpenAI, DeepSeek-R1's open-weight nature allowed researchers to study and build upon the algorithm, though its training data remained private.[90] These reasoning models typically require more computational resources per query compared to traditional LLMs, as they perform more extensive processing to work through problems step-by-step. However, they have shown superior capabilities in domains requiring structured logical thinking, such as mathematics, scientific research, and computer programming.[89]
Efforts to reduce or compensate for hallucinations have employed automated reasoning, RAG (retrieval-augmented generation), fine-tuning, and other methods.[91]
Properties
Scaling laws
Main article: Neural scaling law
The performance of an LLM after pretraining largely depends on the:
cost of pretraining C {\displaystyle C} (the total amount of compute used),
size of the artificial neural network itself, such as number of parameters N {\displaystyle N} (i.e. amount of neurons in its layers, amount of weights between them and biases),
size of its pretraining dataset (i.e. number of tokens in corpus, D {\displaystyle D}).
"Scaling laws" are empirical statistical laws that predict LLM performance based on such factors. One particular scaling law ("Chinchilla scaling") for LLM autoregressively trained for one epoch, with a log-log learning rate schedule, states that:[92] { C = C 0 N D L = A N α + B D β + L 0 {\displaystyle {\begin{cases}C=C_{0}ND\\[6pt]L={\frac {A}{N^{\alpha }}}+{\frac {B}{D^{\beta }}}+L_{0}\end{cases}}} where the variables are
C {\displaystyle C} is the cost of training the model, in FLOPs.
N {\displaystyle N} is the number of parameters in the model.
D {\displaystyle D} is the number of tokens in the training set.
L {\displaystyle L} is the average negative log-likelihood loss per token (nats/token), achieved by the trained LLM on the test dataset.
and the statistical hyper-parameters are
C 0 = 6 {\displaystyle C_{0}=6}, meaning that it costs 6 FLOPs per parameter to train on one token. Note that training cost is much higher than inference cost, where it costs 1 to 2 FLOPs per parameter to infer on one token.[53]
α = 0.34 , β = 0.28 , A = 406.4 , B = 410.7 , L 0 = 1.69 {\displaystyle \alpha =0.34,\beta =0.28,A=406.4,B=410.7,L_{0}=1.69}
Emergent abilities
At point(s) referred to as breaks,[93] the lines change their slopes, appearing on a linear-log plot as a series of linear segments connected by arcs.
Performance of bigger models on various tasks, when plotted on a log-log scale, appears as a linear extrapolation of performance achieved by smaller models. However, this linearity may be punctuated by "break(s)"[93] in the scaling law, where the slope of the line changes abruptly, and where larger models acquire "emergent abilities".[41][94] They arise from the complex interaction of the model's components and are not explicitly programmed or designed.[95]
Furthermore, recent research has demonstrated that AI systems, including large language models, can employ heuristic reasoning akin to human cognition. They balance between exhaustive logical processing and the use of cognitive shortcuts (heuristics), adapting their reasoning strategies to optimize between accuracy and effort. This behavior aligns with principles of resource-rational human cognition, as discussed in classical theories of bounded rationality and dual-process theory.[96]
One of the emergent abilities is in-context learning from example demonstrations.[97] In-context learning is involved in tasks, such as:
reported arithmetics
decoding the International Phonetic Alphabet
unscrambling a word's letters
disambiguating word-in-context datasets[41][98][99]
converting spatial words
cardinal directions (for example, replying "northeast" in response to a 3x3 grid of 8 zeros and a 1 in the top-right), color terms represented in text.[100]
chain-of-thought prompting: In a 2022 research paper, chain-of-thought prompting only improved the performance for models that had at least 62B. Smaller models perform better when prompted to answer immediately, without chain of thought.[101]
identifying offensive content in paragraphs of Hinglish (a combination of Hindi and English), and generating a similar English equivalent of Kiswahili proverbs.[102]
Schaeffer et. al. argue that the emergent abilities are not unpredictably acquired, but predictably acquired according to a smooth scaling law. The authors considered a toy statistical model of an LLM solving multiple-choice questions, and showed that this statistical model, modified to account for other types of tasks, applies to these tasks as well.[103]
Let x {\displaystyle x} be the number of parameter count, and y {\displaystyle y} be the performance of the model.
When y = average Pr ( correct token ) {\displaystyle y={\text{average }}\Pr({\text{correct token}})}, then ( log x , y ) {\displaystyle (\log x,y)} is an exponential curve (before it hits the plateau at one), which looks like emergence.
When y = average log ( Pr ( correct token ) ) {\displaystyle y={\text{average }}\log(\Pr({\text{correct token}}))}, then the ( log x , y ) {\displaystyle (\log x,y)} plot is a straight line (before it hits the plateau at zero), which does not look like emergence.
When y = average Pr ( the most likely token is correct ) {\displaystyle y={\text{average }}\Pr({\text{the most likely token is correct}})}, then ( log x , y ) {\displaystyle (\log x,y)} is a step-function, which looks like emergence.
Interpretation
Large language models by themselves are black boxes, and it is not clear how they can perform linguistic tasks. Similarly, it is unclear if or how LLMs should be viewed as models of the human brain and/or human mind.[104]
Various techniques have been developed to enhance the transparency and interpretability of LLMs. Mechanistic interpretability aims to reverse-engineer LLMs by discovering symbolic algorithms that approximate the inference performed by an LLM. In recent years, sparse coding models such as sparse autoencoders, transcoders, and crosscoders have emerged as promising tools for identifying interpretable features.
Studying a replacement model
Transcoders, which are more interpretable than transformers, have been utilized to develop “replacement models.” In one such study involving the mechanistic interpretation of writing a rhyming poem by an LLM, it was shown that although they are believed to simply predict the next token, they can, in fact, plan ahead.[105]
Explainability
A related concept is AI explainability, which focuses on understanding how an AI model arrives at a given result. Techniques such as partial dependency plots, SHAP (SHapley Additive exPlanations), and feature importance assessments allow researchers to visualize and understand the contributions of various input features to the model's predictions. These methods help ensure that AI models make decisions based on relevant and fair criteria, enhancing trust and accountability.
By integrating these techniques, researchers and practitioners can gain deeper insights into the operations of LLMs, fostering trust and facilitating the responsible deployment of these powerful models.
In another example, the authors trained small transformers on modular arithmetic addition. The resulting models were reverse-engineered, and it turned out they used discrete Fourier transform.[106]
Understanding and intelligence
See also: Philosophy of artificial intelligence and Artificial consciousness
NLP researchers were evenly split when asked, in a 2022 survey, whether (untuned) LLMs "could (ever) understand natural language in some nontrivial sense".[107] Proponents of "LLM understanding" believe that some LLM abilities, such as mathematical reasoning, imply an ability to "understand" certain concepts. A Microsoft team argued in 2023 that GPT-4 "can solve novel and difficult tasks that span mathematics, coding, vision, medicine, law, psychology and more" and that GPT-4 "could reasonably be viewed as an early (yet still incomplete) version of an artificial general intelligence system": "Can one reasonably say that a system that passes exams for software engineering candidates is not really intelligent?"[108][109] Ilya Sutskever argues that predicting the next word sometimes involves reasoning and deep insights, for example if the LLM has to predict the name of the criminal in an unknown detective novel after processing the entire story leading up to the revelation.[110] Some researchers characterize LLMs as "alien intelligence".[111][112] For example, Conjecture CEO Connor Leahy considers untuned LLMs to be like inscrutable alien "Shoggoths", and believes that RLHF tuning creates a "smiling facade" obscuring the inner workings of the LLM: "If you don't push it too far, the smiley face stays on. But then you give it [an unexpected] prompt, and suddenly you see this massive underbelly of insanity, of weird thought processes and clearly non-human understanding."[113][114]
In contrast, some skeptics of LLM understanding believe that existing LLMs are "simply remixing and recombining existing writing",[112] a phenomenon known as stochastic parrot, or they point to the deficits existing LLMs continue to have in prediction skills, reasoning skills, agency, and explainability.[107] For example, GPT-4 has natural deficits in planning and in real-time learning.[109] Generative LLMs have been observed to confidently assert claims of fact which do not seem to be justified by their training data, a phenomenon which has been termed "hallucination".[115] Specifically, hallucinations in the context of LLMs correspond to the generation of text or responses that seem syntactically sound, fluent, and natural but are factually incorrect, nonsensical, or unfaithful to the provided source input.[116] Neuroscientist Terrence Sejnowski has argued that "The diverging opinions of experts on the intelligence of LLMs suggests that our old ideas based on natural intelligence are inadequate".[107]
The matter of LLM's exhibiting intelligence or understanding has two main aspects – the first is how to model thought and language in a computer system, and the second is how to enable the computer system to generate human like language.[107] These aspects of language as a model of cognition have been developed in the field of cognitive linguistics. American linguist George Lakoff presented Neural Theory of Language (NTL)[117] as a computational basis for using language as a model of learning tasks and understanding. The NTL Model outlines how specific neural structures of the human brain shape the nature of thought and language and in turn what are the computational properties of such neural systems that can be applied to model thought and language in a computer system. After a framework for modeling language in a computer systems was established, the focus shifted to establishing frameworks for computer systems to generate language with acceptable grammar. In his 2014 book titled The Language Myth: Why Language Is Not An Instinct, British cognitive linguist and digital communication technologist Vyvyan Evans mapped out the role of probabilistic context-free grammar (PCFG) in enabling NLP to model cognitive patterns and generate human like language.[118][119]
Evaluation
Perplexity
The canonical measure of the performance of an LLM is its perplexity on a given text corpus. Perplexity measures how well a model predicts the contents of a dataset; the higher the likelihood the model assigns to the dataset, the lower the perplexity. In mathematical terms, perplexity is the exponential of the average negative log likelihood per token.
log ( Perplexity ) = − 1 N ∑ i = 1 N log ( Pr ( token i ∣ context for token i ) ) {\displaystyle \log({\text{Perplexity}})=-{\frac {1}{N}}\sum _{i=1}^{N}\log(\Pr({\text{token}}_{i}\mid {\text{context for token}}_{i}))}
Here, N {\displaystyle N} is the number of tokens in the text corpus, and "context for token i {\displaystyle i}" depends on the specific type of LLM. If the LLM is autoregressive, then "context for token i {\displaystyle i}" is the segment of text appearing before token i {\displaystyle i}. If the LLM is masked, then "context for token i {\displaystyle i}" is the segment of text surrounding token i {\displaystyle i}.
Because language models may overfit to training data, models are usually evaluated by their perplexity on a test set.[48] This evaluation is potentially problematic for larger models which, as they are trained on increasingly large corpora of text, are increasingly likely to inadvertently include portions of any given test set.[1]
BPW, BPC, and BPT
In information theory, the concept of entropy is intricately linked to perplexity, a relationship notably established by Claude Shannon.[120] This relationship is mathematically expressed as Entropy = log 2 ( Perplexity ) {\displaystyle {\text{Entropy}}=\log _{2}({\text{Perplexity}})}.
Entropy, in this context, is commonly quantified in terms of bits per word (BPW) or bits per character (BPC), which hinges on whether the language model utilizes word-based or character-based tokenization.
Notably, in the case of larger language models that predominantly employ sub-word tokenization, bits per token (BPT) emerges as a seemingly more appropriate measure. However, due to the variance in tokenization methods across different Large Language Models (LLMs), BPT does not serve as a reliable metric for comparative analysis among diverse models. To convert BPT into BPW, one can multiply it by the average number of tokens per word.
In the evaluation and comparison of language models, cross-entropy is generally the preferred metric over entropy. The underlying principle is that a lower BPW is indicative of a model's enhanced capability for compression. This, in turn, reflects the model's proficiency in making accurate predictions.
Task-specific datasets and benchmarks
A large number of testing datasets and benchmarks have also been developed to evaluate the capabilities of language models on more specific downstream tasks. Tests may be designed to evaluate a variety of capabilities, including general knowledge, bias, commonsense reasoning, and mathematical problem-solving.
One broad category of evaluation dataset is question answering datasets, consisting of pairs of questions and correct answers, for example, ("Have the San Jose Sharks won the Stanley Cup?", "No").[121] A question answering task is considered "open book" if the model's prompt includes text from which the expected answer can be derived (for example, the previous question could be adjoined with some text which includes the sentence "The Sharks have advanced to the Stanley Cup finals once, losing to the Pittsburgh Penguins in 2016."[121]). Otherwise, the task is considered "closed book", and the model must draw on knowledge retained during training.[122] Some examples of commonly used question answering datasets include TruthfulQA, Web Questions, TriviaQA, and SQuAD.[122]
Evaluation datasets may also take the form of text completion, having the model select the most likely word or sentence to complete a prompt, for example: "Alice was friends with Bob. Alice went to visit her friend, ____".[1]
Some composite benchmarks have also been developed which combine a diversity of different evaluation datasets and tasks. Examples include GLUE, SuperGLUE, MMLU, BIG-bench, HELM, and HLE (Humanity's Last Exam).[120][122] OpenAI has released tools for running composite benchmarks, but noted that the eval results are sensitive to the prompting method.[123][124] Some public datasets contain questions that are mislabeled, ambiguous, unanswerable, or otherwise of low-quality, which can be cleaned to give more reliable benchmark scores.[125]
Bias in LLMs may be measured through benchmarks such as CrowS-Pairs (Crowdsourced Stereotype Pairs),[126] Stereo Set,[127] and the more recent Parity Benchmark.[128] Additionally, fact-checking and misinformation detection are becoming increasingly crucial evaluation areas for LLMs. A recent study by Caramancion (2023) compared the fact-checking accuracy of prominent LLMs—including OpenAI’s ChatGPT 3.5 and 4.0, Google’s Bard, and Microsoft’s Bing AI—against independent fact-checking agencies such as PolitiFact and Snopes. The results demonstrated a moderate proficiency in fact verification, with GPT-4 achieving the highest accuracy at 71%, but still lagging behind human fact-checkers in contextual comprehension and nuanced reasoning. This underscores the evolving but incomplete ability of LLMs to discern fact from deception, highlighting the need for continued advancements in AI-driven fact-checking methodologies.[129]
It was previously standard to report results on a heldout portion of an evaluation dataset after doing supervised fine-tuning on the remainder. It is now more common to evaluate a pre-trained model directly through prompting techniques, though researchers vary in the details of how they formulate prompts for particular tasks, particularly with respect to how many examples of solved tasks are adjoined to the prompt (i.e. the value of n in n-shot prompting).
Adversarially constructed evaluations
Because of the rapid pace of improvement of large language models, evaluation benchmarks have suffered from short lifespans, with state of the art models quickly "saturating" existing benchmarks, exceeding the performance of human annotators, leading to efforts to replace or augment the benchmark with more challenging tasks.[130] In addition, there are cases of "shortcut learning" wherein AIs sometimes "cheat" on multiple-choice tests by using statistical correlations in superficial test question wording in order to guess the correct responses, without necessarily understanding the actual question being asked.[107]
Some datasets have been constructed adversarially, focusing on particular problems on which extant language models seem to have unusually poor performance compared to humans. One example is the TruthfulQA dataset, a question answering dataset consisting of 817 questions which language models are susceptible to answering incorrectly by mimicking falsehoods to which they were repeatedly exposed during training. For example, an LLM may answer "No" to the question "Can you teach an old dog new tricks?" because of its exposure to the English idiom you can't teach an old dog new tricks, even though this is not literally true.[131]
Another example of an adversarial evaluation dataset is Swag and its successor, HellaSwag, collections of problems in which one of multiple options must be selected to complete a text passage. The incorrect completions were generated by sampling from a language model and filtering with a set of classifiers. The resulting problems are trivial for humans but at the time the datasets were created state of the art language models had poor accuracy on them. For example:
We see a fitness center sign. We then see a man talking to the camera and sitting and laying on a exercise ball. The man...
a) demonstrates how to increase efficient exercise work by running up and down balls.
b) moves all his arms and legs and builds up a lot of muscle.
c) then plays the ball and we see a graphics and hedge trimming demonstration.
d) performs sit ups while on the ball and talking.[132]
BERT selects b) as the most likely completion, though the correct answer is d).[132]
Limitations of LLM benchmarks
Benchmarks can become outdated rapidly. Once a model attains near-perfect scores on a given benchmark, that benchmark ceases to serve as a meaningful indicator of progress. This phenomenon, known as "benchmark saturation," necessitates the development of more challenging and nuanced tasks to continue advancing LLM capabilities. For instance, traditional benchmarks like HellaSwag and MMLU have seen models achieving high accuracy already.
Wider impact
In 2023, Nature Biomedical Engineering wrote that "it is no longer possible to accurately distinguish" human-written text from text created by large language models, and that "It is all but certain that general-purpose large language models will rapidly proliferate... It is a rather safe bet that they will change many industries over time."[133] Goldman Sachs suggested in 2023 that generative language AI could increase global GDP by 7% in the next ten years, and could expose to automation 300 million jobs globally.[134][135] Brinkmann et al. (2023)[136] also argue that LLMs are transforming processes of cultural evolution by shaping processes of variation, transmission, and selection.
Memorization and copyright
Further information: Artificial intelligence and copyright
Memorization is an emergent behavior in LLMs in which long strings of text are occasionally output verbatim from training data, contrary to typical behavior of traditional artificial neural nets. Evaluations of controlled LLM output measure the amount memorized from training data (focused on GPT-2-series models) as variously over 1% for exact duplicates[137] or up to about 7%.[138]
A 2023 study showed that when ChatGPT 3.5 turbo was prompted to repeat the same word indefinitely, after a few hundreds of repetitions, it would start outputting excerpts from its training data.[139]
Security
Some commenters expressed concern over accidental or deliberate creation of misinformation, or other forms of misuse.[140] For example, the availability of large language models could reduce the skill-level required to commit bioterrorism; biosecurity researcher Kevin Esvelt has suggested that LLM creators should exclude from their training data papers on creating or enhancing pathogens.[141]
The potential presence of "sleeper agents" within LLMs is another emerging security concern. These are hidden functionalities built into the model that remain dormant until triggered by a specific event or condition. Upon activation, the LLM deviates from its expected behavior to make insecure actions.[142]
LLM applications accessible to the public, like ChatGPT or Claude, typically incorporate safety measures designed to filter out harmful content. However, implementing these controls effectively has proven challenging. For instance, a 2023 study[143] proposed a method for circumventing LLM safety systems. In 2025, The American Sunlight Project, a non-profit, published a study[144] showing evidence that the so-called Pravda network, a pro-Russia propaganda aggregator, was strategically placing web content through mass publication and duplication with the intention of biasing LLM outputs. The American Sunlight Project coined this technique "LLM grooming," and pointed to it as a new tool of weaponizing AI to spread disinformation and harmful content.[144][145] Similarly, Yongge Wang[146] illustrated in 2024 how a potential criminal could potentially bypass ChatGPT 4o's safety controls to obtain information on establishing a drug trafficking operation. External filters, circuit breakers and overrides have been posed as solutions.[citation needed]
Algorithmic bias
Main article: Algorithmic bias
While LLMs have shown remarkable capabilities in generating human-like text, they are susceptible to inheriting and amplifying biases present in their training data. This can manifest in skewed representations or unfair treatment of different demographics, such as those based on race, gender, language, and cultural groups.[147] Since English data is overrepresented in current large language models' training data, it may also downplay non-English views.[148]
Stereotyping
AI models can reinforce a wide range of stereotypes, including those based on gender, ethnicity, age, nationality, religion, or occupation. This can lead to outputs that homogenize, or unfairly generalize or caricature groups of people, sometimes in harmful or derogatory ways.[149][150]
Notably, gender bias refers to the tendency of these models to produce outputs that are unfairly prejudiced towards one gender over another. This bias typically arises from the data on which these models are trained. Large language models often assign roles and characteristics based on traditional gender norms.[147] For example, it might associate nurses or secretaries predominantly with women and engineers or CEOs with men.[151]
Selection bias
Selection bias refers the inherent tendency of large language models to favor certain option identifiers irrespective of the actual content of the options. This bias primarily stems from token bias—that is, the model assigns a higher a priori probability to specific answer tokens (such as “A”) when generating responses. As a result, when the ordering of options is altered (for example, by systematically moving the correct answer to different positions), the model’s performance can fluctuate significantly. This phenomenon undermines the reliability of large language models in multiple-choice settings.[152][153]
Political bias
Political bias refers to the tendency of algorithms to systematically favor certain political viewpoints, ideologies, or outcomes over others. Language models may also exhibit political biases. Since the training data includes a wide range of political opinions and coverage, the models might generate responses that lean towards particular political ideologies or viewpoints, depending on the prevalence of those views in the data.[154]
Energy demands
The energy demands of LLMs have grown along with their size and capabilities. Data centers that enable LLM training require substantial amounts of electricity. Much of that electricity is generated by non-renewable resources that create greenhouse gases and contribute to climate change.[155] Nuclear power and geothermal energy are two options tech companies are exploring to meet the sizable energy demands of LLM training.[156] The significant expense of investing in geothermal solutions has led to major shale producers like Chevron and Exxon Mobil advocating for tech companies to use electricity produced via natural gas to fuel their large energy demands.[157]
The Alan MacMasters hoax was a hoax that appeared on the English Wikipedia for more than ten years. In February 2012, a group of British students edited the encyclopedia's article about electric toasters and inserted the false claim that a man named Alan MacMasters invented the toaster in 1893. One of the students created a separate article about the fictitious Alan MacMasters in February 2013 and embellished it with further details in the following years. The fake article was cited by several newspapers and organizations until the hoax was exposed in July 2022.
The actual development of the pop-up toaster was based on technologies and features invented between 1890 and 1920 by various people and companies.
Origins
On 6 February 2012, University of Surrey aerospace engineering student Alan MacMasters was at a university lecture on dynamics where the class was warned not to use Wikipedia as a source. Additionally, the lecturer pointed out that his friend, named Maddy Kennedy, had edited the Wikipedia article about toasters, falsely claiming he was the inventor.[1][2][3][4]
After the lecture, Alan and his friends visited the toaster article on Wikipedia, where one of his friends, Alex, edited the article to replace the lecturer's friend's name with Alan MacMasters, claiming he invented the toaster in Edinburgh, Scotland, in 1893.[1][2][4][5][6]
A year later, Alex contemplated the extent to which he could escalate the prank. In February 2013, he created an article dedicated to Alan MacMasters, including an image of himself manipulated to resemble a 19th-century photograph, and published it on Wikipedia. Alex and other editors extended and embellished the fictitious biography over time.[1][2][6]
In the article, Alex mentioned that the product was not commercially successful.[5] He also attributed the invention of the electric kettle to MacMasters and suggested that the toaster had contributed to one of Britain's earliest fatal appliance fires.[2][5] One fabricated anecdote recounted a woman whose kitchen table caught fire after the toaster's heating elements melted.[2][5] Another falsehood he added was that MacMasters had assisted in developing lighting systems for the London Underground.[1]
Impact
While the article had started out as a jest, many people and organizations accepted its claims as fact and perpetuated the false story.[1][3][2] This created a case of circular reporting, as Alex and others then used these sources citing MacMasters as the inventor of the toaster to prop up the false information on Wikipedia.[3]
Others repeating the story included newspapers such as The Scotsman and The Mirror,[2] the Chicago History Museum,[7] Purdue University,[8] and the Hagley Museum and Library in Delaware.[1][9]
More than twelve books in multiple languages named MacMasters as the inventor.[1] A primary school in Scotland dedicated a day to MacMasters.[1] In a response to a request for nominations from the Bank of England, MacMasters was nominated to appear on a £50 note, and was preselected as one of the 989 eligible names out of 227,299 nominations.[1][10] During the 2014 Scottish independence referendum, Scottish Government-funded organizations cited Alan's story as evidence of how an independent Scotland could succeed.[2][11][12] During an appearance on the BBC cooking show Great British Menu, chef Scott Smith created a dessert in honor of MacMasters.[1][4]
Discovery and aftermath
In July 2022, a Kent-based teenage student named Adam became suspicious of the photograph on Alan MacMasters' Wikipedia page and, upon scrutinizing it further, discovered that it was edited and not legitimate.[1][4] Adam subsequently posted his findings to Reddit.[1] This research was prompted after his teacher spoke about MacMasters in class and Adam looked up the article of the supposed inventor.[6] However, Adam was unaware that the entire article was a hoax.[1] A viewer of the Reddit post reported their concern on the Internet forum Wikipediocracy, where users discovered the article's fraudulent nature.[1] Soon after this, the page was labeled as a hoax and marked for deletion.[1] Alex's Wikipedia account, which he used to perpetrate the hoax, was subsequently blocked from the platform.[1][4]
In an interview published on Wikipediocracy in 2022, the creator of the hoax said that he initially thought the prank would not cause much harm, that awareness that the article was a hoax had actually been "widespread", and that many subsequent embellishments to the article had been made by other editors.[11] He described the first time he realized the prank was harmful was when he read a book about Victorian inventors and found Alan MacMasters listed as one of the inventors.[11]
The Republic of Ireland Act 1948[a] (No. 22 of 1948) is an Act of the Oireachtas which declares that the description of Ireland is the Republic of Ireland, and vests in the president of Ireland the power to exercise the executive authority of the state in its external relations, on the advice of the Government of Ireland. The Act was signed into law on 21 December 1948 and came into force on 18 April 1949, Easter Monday,[1][2] the 33rd anniversary of the beginning of the Easter Rising.
The Act ended the remaining statutory role of the British monarchy in relation to Ireland, by repealing the 1936 External Relations Act, which had vested in George VI, in his capacity as a symbol of the cooperation of the nations that were members of the Commonwealth with which Ireland associated itself, and his successors those functions which the Act now transferred to the President.
Text of the Act
The Republic of Ireland Act consists of five brief sections, set out in full as follows:
Number 22 of 1948
The Republic of Ireland Act, 1948
An Act to repeal the Executive Authority (External Relations) Act, 1936, to declare that the description of the State shall be the Republic of Ireland, and to enable the President to exercise the executive power or any executive function of the state in or in connection with its external relations. (21 December 1948)
Be it enacted by the Oireachtas as follows:—
1.—The Executive Authority (External Relations) Act, 1936 (No. 58 of 1936), is hereby repealed.
2.—It is hereby declared that the description of the State shall be the Republic of Ireland.
3.—The President, on the authority and on the advice of the Government, may exercise the executive power or any executive function of the State in or in connection with its external relations.
4.—This Act shall come into operation on such day as the Government may by order appoint.
5.—This Act may be cited as The Republic of Ireland Act, 1948.
British monarch
Section 1 of the Act repealed the Executive Authority (External Relations) Act 1936. By doing so the Act abolished the last remaining functions of the British monarch (then King George VI) in relation to the Irish state. These functions had related to the issuance and acceptance of letters of credence of diplomatic and consular representatives and the conclusion of international agreements. Section 3 provides that the President of Ireland may instead exercise these functions and any other functions in relation to the state's external (or foreign) relations.
The Commonwealth
At the time the Act came into force, John A. Costello, the Taoiseach whose government introduced the Act, believed that Ireland did not have a King and had not been a member of the Commonwealth since 1936.[3] His government's view was that Ireland was already a republic and that the Act would not create a republic but rather achieve a "clarification of [Ireland's] constitutional status."[4] These views were shared by the Irish opposition leader of the time, Éamon de Valera.[5] Indeed, Irish leaders had on several previous occasions declared that Ireland was a republic and not a Commonwealth member, but that it was associated with the Commonwealth.[6]
The Irish view of things was not shared by the other members of the Commonwealth. Until Ireland brought the Act into force, it was still regarded by the members as forming part of "His Majesty’s dominions". When Ireland adopted its 1937 Constitution, which made no reference to the King, the United Kingdom Government announced that it and the other Commonwealth Governments were "[still] prepared to treat … Ireland, as a member of the British Commonwealth of Nations".[7] After all, in their view, the King was still empowered by Ireland to fulfill certain functions as Ireland's statutory agent under the External Relations Act 1936. With that Irish Act now being repealed, there was no longer any basis, however tenuous, to consider Ireland as continuing to have a King or to be part of His Majesty’s dominions and therefore within the Commonwealth. In their view, Ireland had now declared itself a republic for the first time bringing its membership of the Commonwealth to an end. Ironically, the Taoiseach chose to announce the repeal of the External Relations Act while on an official visit to Canada, the same country whose constitutional status had been the basis for the establishment of the Irish Free State.
The London Declaration, which permitted republics to remain within the Commonwealth, was made shortly afterwards in response to India's desire to continue as a member once its new republican constitution was finalised. However, the Irish government did not reapply for membership of the Commonwealth. De Valera was opposed to this and considered applying for membership after his return to office in the 1950s.[8]
Republic of Ireland description
Main article: Names of the Irish state
Section 2 of the Act provides:
It is hereby declared that the description of the State shall be the Republic of Ireland.
The Act did not declare that Ireland a republic, nor did it change the official name of the state which continued to be Éire (in Irish) and Ireland (in English) as prescribed in the Constitution. The Act provided for a description for the State. The distinction between a description and a name has sometimes caused confusion. Costello explained the difference in the following way:[9]
If I say that my name is Costello and that my description is that of senior counsel, I think that will be clear to anybody who wants to know. If the Senator [Helena Concannon] will look at Article 4 of the Constitution she will find that the name of the State is Éire. Section 2 of this Bill declares that "this State shall be described as the Republic of Ireland." Its name in Irish is Éire and in the English language, Ireland. Its description in the English language is "the Republic of Ireland."
Background
See also: Irish head of state from 1922 to 1949
In 1945, when asked if he planned to declare a republic, the Taoiseach Éamon de Valera had replied, "we are a republic",[10] which he had not said in the previous eight years. He also insisted that Ireland had no king, but simply used an external king as an "organ" in international affairs.
In October 1947, de Valera asked the attorney-general, Cearbhall Ó Dálaigh, to draft a bill to repeal the External Relations Act,[11] and by 1948 a draft of the bill included a reference to the state as being a republic.[12] In the end, the draft bill was never submitted to the Oireachtas for approval.
By the eve of the 1948 Irish general election the United Kingdom's Representative to Ireland, Lord Rugby, reported that the annulment of the External Relations Act was inevitable. He remarked 'No party has left the door open to any other course'.[13] The result of the election saw a new Irish government formed under the leadership of John A. Costello.
Costello made the announcement that a bill to repeal the External Relations Act was to be introduced when he was in Ottawa, during an official visit to Canada. David McCullagh has suggested that it was a spur of the moment reaction to offence caused by the Governor General of Canada,[14] Lord Alexander, who was of Northern Ireland descent, who allegedly placed loyalist symbols, notably a replica of the famous Roaring Meg cannon used in the Siege of Derry, before an affronted Costello at a state dinner. What is certain is that an agreement that there would be separate toasts for the King and for the President of Ireland was broken.[14] The Irish position was that a toast to the King, instead of representing both countries, would not include Ireland. Only a toast to the King was proposed, to the fury of the Irish delegation.[14] Shortly afterwards Costello announced the plan to repeal the External Relations Act.
However, according to all but one of the ministers in Costello's cabinet, the decision to repeal the External Relations Act had already been made before Costello's Canadian visit.[15] Costello's revelation of the decision was because the Sunday Independent (an Irish newspaper) had discovered the fact and was about to "break" the story as an exclusive. Nevertheless, one minister, Noel Browne, gave a different account in his autobiography, Against the Tide. He claimed Costello's announcement was done in a fit of anger of his treatment by the Governor General and that when he returned, Costello, at an assembly of ministers in his home, offered to resign because of his manufacture of a major government policy initiative on the spot in Canada. Yet according to Browne, all the ministers agreed that they would refuse to accept the resignation and also agreed to manufacture the story of a prior cabinet decision.[16]
The evidence of what really happened remains ambiguous. There is no record of a prior decision to repeal the External Relations Act before Costello's Canadian trip, among cabinet papers for 1948, which supports Browne's claim.[15] However, the Costello government refused to allow the Secretary to the Government, Maurice Moynihan, to attend cabinet meetings and take minutes, because they believed he was too close to the opposition leader, Éamon de Valera.[17] Rather than entrust the minute-taking to Moynihan, the cabinet entrusted it to a Parliamentary Secretary (junior minister), Liam Cosgrave. Given that Cosgrave had never kept minutes before, his minutes, at least early on in the government, proved to be only a limited record of government decisions. So whether the issue was never raised, was raised but undecided on, was subjected to a decision taken informally, or was subjected to a decision taken formally, remains obscure on the basis of the 1948 cabinet documentation.[15]
Introduction of the bill
The Republic of Ireland Bill was introduced in 1948 by the new Taoiseach, John A. Costello of Fine Gael.
The Act was enacted with all parties voting for it. De Valera did suggest that it would have been better to reserve the declaration of the republic until Irish unity had been achieved, a comment hard to reconcile with his 1945 claim that the Irish state was already a republic.
Response
United Kingdom
The United Kingdom responded to the Republic of Ireland Act by enacting the Ireland Act 1949. This Act formally asserted that the Irish state had, when the Republic of Ireland Act came into force, ceased "to be part of His Majesty's dominions"[18] and accordingly was no longer within the Commonwealth. Nonetheless, the United Kingdom statute provided that Irish citizens would not be treated as aliens under British nationality law. This, in effect, granted them a status similar to the citizens of Commonwealth countries.[19]
Between the enactment of the Constitution of Ireland in 1937 and the enactment of the Ireland Act 1949, the United Kingdom had formally decided upon the (anglicised) "Eire" as its name for the Irish state. The 1949 Act now provided that "the part of Ireland heretofore known as Eire" could be referred to in future UK legislation as the "Republic of Ireland".[20] The UK's continued aversion to using "Ireland" as the formal name for the state due to the fact it did not (and does not) comprise the entirety of the island of the same name remained a source of diplomatic friction for several decades afterwards.
The UK's Ireland Act also gave a legislative guarantee that Northern Ireland would continue to remain a part of the United Kingdom unless the Parliament of Northern Ireland formally expressed a wish to join a United Ireland; this "unionist veto" proved to be controversial during the Act's passage through Westminster, as well as in the Irish state and amongst Northern Ireland's nationalist community. The guarantee was replaced in 1973, when the Parliament of Northern Ireland was abolished, by a new guarantee based on "the consent of the majority of the people of Northern Ireland".[21]
On the day the Act came into force, 18 April 1949, King George VI sent the following message to the President of Ireland, Seán T. O'Kelly:[22]
I send you my sincere good wishes on this day, being well aware of the neighbourly links which hold the people of the Republic of Ireland in close association with my subjects of the United Kingdom. I hold in most grateful memory the services and sacrifices of the men and women of your country who rendered gallant assistance to our cause in the recent war and who made a notable contribution to our victories. I pray that every blessing may be with you today and in the future.
— GEORGE R.
Irish peers
From the Acts of Union 1800, when the UK House of Lords noted someone's succession to an Irish peerage, the Clerk of the Parliaments informed the Clerk of the Crown in Ireland in Dublin to update the electoral register for Irish representative peer. Such elections ceased in 1922 and the office of Clerk of the Crown was formally abolished in 1926, when the last holder, Gerald Horan, became first Master of the High Court. Nevertheless, the Clerk of the Parliaments continued to inform Horan in the old manner until the Irish government, reviewing administration for the commencement of the Republic of Ireland Act, informed the Lords in late 1948 that the Clerk of the Crown in Ireland no longer existed.[23]
Church of Ireland
The Book of Common Prayer of the all-island Church of Ireland was modelled on that of the Church of England and included three "state prayers": for "our most gracious Sovereign Lord, King George", the royal family, and the Commonwealth. The church was historically associated with the Protestant Ascendancy and had been the established church until 1871; its "southern" membership (one-third of the total) was mostly unionist before 1922 and pro-British thereafter. In late 1948, archbishops John Gregg and Arthur Barton devised replacement prayers to be used in the republic, at first temporarily until the 1949 general synod would update the Book of Common Prayer. A grassroots campaign led by Hugh Maude of Clondalkin opposed any change, and the 1950 synod authorised a compromise, whereby the old prayers remained in Northern Ireland, and the republic used a "Prayer for the President and all in authority" and "A Prayer for King George the Sixth … in whose dominions we are not accounted strangers" (an allusion to the Ireland Act 1949). Likewise, the liturgy for morning and evening prayers includes "O Lord, save the Queen" in Northern Ireland and "O Lord, guide and defend our rulers" in the republic.[24][25] Miriam Moffitt notes that Maude's supporters were mostly older church members.[24]
Reassessment
In 1996, the Constitution Review Group reviewed the full text of the Constitution. It considered whether the name of the state should be amended to declare that Ireland should be named "Republic of Ireland". It decided against recommending such an amendment.[26] This was the second time that such an amendment was considered by committee, which considered every provision of the constitution.
Lady Blue is an American detective and action-adventure television series on the American Broadcasting Company (ABC) network. Created by Robert Vincent O'Neil and produced by David Gerber, the short-lived series was canceled by ABC after 13 episodes, which aired from September 15, 1985, to January 25, 1986. It had been picked up to series after ABC aired a two-hour television film pilot on April 15, 1985. The show revolves around Chicago detective Katy Mahoney (Jamie Rose) and her violent methods of handling cases. The supporting cast includes Danny Aiello, Ron Dean, Diane Dorsey, Bruce A. Young, Nan Woods, and Ricardo Gutierrez. Johnny Depp also guest-starred on the series in one of his earliest roles. With cinematography by Jack Priestley, the episodes were filmed on location in Chicago. Television critics noted Lady Blue's emphasis on violence, calling Mahoney "Dirty Harriet" (after Clint Eastwood's character Dirty Harry). Rose said she joined the project after being drawn to its genre. She prepared for the role by watching Eastwood's films, received advice from Eastwood on how to handle a gun, and practiced at a shooting range.
After the pilot aired, Lady Blue was criticized by several watchdog organizations (particularly the National Coalition on Television Violence) as the most violent show on television. After the fourth episode, ABC placed it on hiatus for a month, moved the series from Thursdays to Saturdays, then canceled it in early 1986, partially due to the complaints about excessive violence. Critical reception to the series was primarily negative during its run, but television studies author Cary O'Dell questions whether that stemmed from contemporary sexism. The series' rights are owned by Metro-Goldwyn-Mayer, which has not released Lady Blue on DVD, Blu-ray, nor any online streaming service.
Premise and characters
An image of a woman with bright red, curly hair. She is holding a gun and looking toward the camera.
Critics compared Katy Mahoney to Dirty Harry due to her frequent use of violence.[1][2][3] Jamie Rose was advised by Clint Eastwood on how to handle a gun, and practiced at a shooting range.[4]
A detective and action-adventure television series,[1][5] Lady Blue revolves around Chicago investigator Katy Mahoney (Jamie Rose), her violent means of dealing with criminals and tension with her co-workers.[1] She works in the Violent Crimes Division of the Chicago Police Department.[2] The New York Observer's Bryan Reesman described Mahoney as "the fiery red head" with a "trigger happy" personality and "violent excesses".[1] She frequently uses a .357 Magnum (which John J. O'Connor of The New York Times called "a grotesque extension of her right arm"),[5] and was introduced as capable of "read[ing] a crime in progress like most guys read the sports page".[2][6]
Mahoney's reliance on violence is emphasized in the opening scene of the pilot; she sees a bank robbery while she is in a beauty parlor, shoots and kills three of the perpetrators, and returns to the salon for a pedicure.[1][5] Television critics and the show's promotional materials called Mahoney "Dirty Harriet" and "Dirty Harriette", comparing her aggressive behavior to Clint Eastwood's character Dirty Harry,[1][3][4] and Jon Anderson of the Chicago Tribune described her as "somewhat like Quick Draw McGraw with touches of John Wayne and Clint Eastwood".[7] According to Rose, Mahoney was inspired by Dirty Harry, Wayne, and Rambo.[7] Mahoney and other characters refer to the number of excessive-force complaints filed against her during the series,[5] and she often has difficulties with Internal Affairs.[8]
Although Mahoney was portrayed at odds with most of her superiors, her boss Lt. Terry McNichols (Danny Aiello) is more sympathetic and understanding towards her.[1] McNichols is portrayed as fond of chili dogs and appreciative of Mahoney's more unorthodox methods of handling criminals, although he still criticizes her reliance on violence.[5][9] Rose described McNichols as similar to a character in the crime drama The Sopranos.[1] Describing Aiello's performance, O'Connor wrote that McNichols "offer[ed] an uncanny impersonation of the punch-drunk Slapsie Maxie Rosenbloom in a 1940's movie".[5]
Mahoney's father, brother, and married lover were killed in the line of duty before the series begins, and O'Connor connected these events to the character's "toughness and determination to survive".[5] Other characters include detective Gino Gianelli (Ron Dean) and his wife Rose (Diane Dorsey), Officer Cassidy (Bruce A. Young), McNichols' niece Willow (Nan Woods), and Mahoney's informant Harvey (Ricardo Gutierrez).[2] In one of his earliest roles, American actor Johnny Depp guest-starred in an episode as the brother of a serial killer.[10][11] Mexican actress Katy Jurado appeared in the pilot as cocaine kingpin Dona Maria Theresa,[8] and American actors Ajay Naidu and Jim Brown portrayed "worldly-wise waif" Paquito and a "South Side drug czar", respectively.[12] Aiello's best friend was an extra in the series, the cast and crew calling his character "Detective Joe Background".[1] Tom Shales of The Washington Post described the show's tone as "baldly campy [and] ultra-violent".[13]
Production
A black-and-white photograph of a neighborhood; it is taken from an aerial point of view.
One episode was filmed in the Cabrini–Green Homes, pictured in 1999.
Lady Blue was created by Robert Vincent O'Neil.[14] The executive producer was David Gerber.[10] Directors Guy Magar and Gary Nelson worked on the series,[10][15] while Jack Priestley was the cinematographer.[15] Produced by MGM Television and David Gerber Productions,[4][10][15] its musical score was composed by John Cacavas.[15][16] Actress Arnetia Walker performed the show's theme song, "Back to the Blue".[2][9] Lady Blue was filmed on location in various areas of Chicago,[4][10] including the Cabrini–Green Homes.[1] Rose recalled having a difficult time in Cabrini Green since the residents threatened the cast and crew and threw bottles at them during filming.[1]
Mahoney was Rose's first role after playing Vickie Gioberti in the soap opera Falcon Crest; Reesman wrote that the decision to cast Rose in Lady Blue was a surprise, since she was primarily known for appearing as a child with Bugs Bunny in a Kool-Aid commercial. According to Reesman, Mahoney's "steely nerve and conservative stance on crime" contrasted with Rose's "more upbeat, fun-loving, liberal persona". Rose said that she was drawn to the show's genre: "Action shows are so fun because I got to be strapped to things, hoisted over things, shoot the gun, and jump on moving cars. It was like doing a western."[1] According to the Orange County Register, Mahoney is one of the actress's best-known roles.[17]
To prepare for Lady Blue, Rose watched Clint Eastwood films (including the Dirty Harry franchise) and practiced steadying her gun hand.[4][7] She had worked with Eastwood in the 1984 film Tightrope and a portion of the anthology series Amazing Stories, and received advice on how to mimic using a gun from Eastwood.[4] In addition to Eastwood's assistance, Rose practiced gun-handling at a Chicago shooting range.[4] Although Rose described her role as "physically demanding", she said she was not attempting method acting and relied on stunt doubles during filming.[7] Rose resisted comparisons to Dirty Harry, and said: "It's still going to be a lot different because I'm a woman and I can show lots more emotions than Mr. Eastwood."[18]
According to Jamie Rose, Lady Blue had a similar concept as the crime dramas Police Woman and Get Christie Love!; Reesman stated that the latter was not as violent as Lady Blue.[1] John J. O'Connor compared the series' violence to Eastwood's work, and saw it as a combination of Wonder Woman and Dick Tracy comic strips.[5] In the 2011 book Triumph of the Walking Dead: Robert Kirkman's Zombie Epic on Page and Screen, horror fiction writer Vince A. Liaguno described Lady Blue and NYPD Blue as part of a movement towards "grittier depictions of violence".[19] In a 2017 interview, Rose said that Lady Blue was the most violent series of its time and there had been little public exposure to a character as "bloodthirsty" as Mahoney; however, she said that the series was less graphic than future television programs.[1]
Episodes
No. Title Directed by Written by Original release date
1 "Pilot" Gary Nelson Robert Vincent O'Neil April 15, 1985[8]
Homicide detective Katy Mahoney is transferred to the "Matron Squad" of the Chicago Police Department after several charges of excessive force are filed against her. While investigating a shoplifting case and the murders of two women and their children, she discovers that they are all related to cocaine trafficking.
2 "Death Valley Day" Virgil Vogel Nancy Audley and Howard Chesley September 26, 1985[8]
Mahoney investigates a murder at a housing project and discovers that the area is terrorized by Alvin Banger and his gang. When she learns that Banger forces children to steal from stores and homes, she decides to bring down the gang.
3 "Romeo and Juliet" Robert Vincent O'Neil Mark Rodgers October 3, 1985[20]
During a war between two rival street gangs, a man and a woman from the opposing sides develop feelings for one another and Mahoney tries to help them find a future together.
4 "Beasts of Prey" Guy Magar Anthony Lawrence and Nancy Lawrence October 10, 1985[20]
Mahoney tracks a serial killer while investigating a string of South Side murders. She asks her informer Della to do some undercover work. Mahoney is devastated when Della's mutilated body is found but Della managed to capture the identity of the killers on film before she died.
5 "The Widow-Maker" Mike Vejar Allison Hock October 17, 1985[20]
Mahoney searches for a Vietnamese assassin who is programmed to kill former soldiers and refugees who have moved to America.
6 "The Hunter" Christian I. Nyby II Robert Vincent O'Neil November 16, 1985[20]
Mahoney investigates a series of murders committed with unconventional weapons and poisons.
7 "Portrait of Death" Mike Vejar Anthony Lawrence and Nancy Lawrence November 23, 1985[20]
An imprisoned criminal mastermind hires hitmen to kill everyone responsible for her conviction, starting with Mahoney and her former lover. Mahoney discovers that her ex-partner is a lawyer with questionable connections.
8 "Terror" John Florea Bill Driskill November 30, 1985[20]
While infiltrating a terrorist organization, Mahoney discovers that their main objective is to dismantle Chicago's political system.
9 "Designer White" Arnold Laven Michael Ahnemann December 7, 1985[20]
Hot on the trail of a drug ring, Mahoney becomes a victim when she is injected with a new type of bootleg hallucinogen designer drug during an encounter with a drug dealer in a police raid. Plagued by these drug-induced hallucinations, strange visions, paranoia and paralysis, Katy flees by wandering around the city and McNichols finds himself in a race against time to find Katy before she hurts someone.
10 "Death Grip" Guy Magar Mark Rodgers December 21, 1985[20]
While tracking down a hitman who kills local drug kingpins, Mahoney discovers that he is part of a larger plot to establish an international narcotics operation.
11 "Scorpio's Sting" John Hancock Robert Vincent O'Neill January 11, 1986[20]
Mahoney is tasked to find a former Green Beret and his gang, who are killing people for thrills.
12 "Sylvie" Mike Vejar Michael Ahemann January 18, 1986[20]
During an investigation, Mahoney discovers that a policewoman was murdered to cover up a scandal involving politicians, pornographers, and bankers.
13 "Maximum Force" Jerry Jameson Mark Rodgers January 25, 1986[8]
Mahoney and detective Gino Gianelli are kidnapped by a group seeking vengeance for an arrest.
14 "Willow's Cowboy" Jerry Jameson Mark Rodgers February 1, 1986[21]
While searching for Terry McNichols' missing niece, Mahoney becomes involved with a group of cowboys trying to steal a shipment of bull semen.
Broadcast history
Thirteen episodes of Lady Blue were broadcast on ABC between September 15, 1985 and January 25, 1986.[1][8] The pilot episode was aired as a television film on April 15, 1985,[8] before it was aired as part of the series in September of that year.[1][4] According to Lee Margulies of the Los Angeles Times, the pilot film received high ratings.[22] When the series began, its emphasis on violence was criticized[1][7] and it was included on watchdog organization lists.[5] A total of 18 characters were killed in the pilot, and producers had promised future episodes would feature more deaths.[23] The National Coalition on Television Violence called it the "most violent program" on television during the series' run.[24] In response to the criticism, Rose said that Lady Blue was set in "more of the heroic fantasy world" and compared Mahoney to a superhero; she explained that series was not intended to be a realistic representation of the police.[25]
Lady Blue was initially broadcast on Thursday nights at 9 pm EST;[26] the series ranked third in its time slot, behind the half-hour sitcoms Cheers and Night Court and the detective series Simon & Simon.[7] After four episodes aired, it was moved to Saturday nights at 9 pm EST to accommodate The Colbys and to replace Lime Street, which was forced to cease production after the death of Samantha Smith and her father in a plane crash during its production.[7][22] ABC announced that it ordered a limited number of episodes of Lady Blue in its new time, but the series would be moved to another day "without interrupting the weekly flow" if it was successful.[22] While on that Saturdays-at-9-pm timeslot, it aired against The Golden Girls and 227,[5][27] and continued to receive complaints of excessive violence.[27]
ABC canceled Lady Blue in 1986.[2][28] Reesman also attributed the decision to low ratings.[1] After the end of the series, Rose said: "It was still a great experience. You don't get much opportunity to star in your own series, especially if you're a woman."[28] Lady Blue was rebroadcast on Lifetime, following the network's tradition of airing shows depicting female characters in traditionally-male occupations; other examples include female private detectives in Veronica Clare and Partners in Crime as well as a female physician in Kay O'Brien.[29] The series has not been released on DVD, Blu-ray or an online-streaming service.[1][30] Metro-Goldwyn-Mayer owns the rights to Lady Blue, but a studio spokesperson said that there were no plans for a home release.[1]
Critical reception
During its run, Lady Blue received primarily negative reviews due to its emphasis on violence.[8][24] Although O'Connor criticized the series for its "mindless violence and questionable law enforcement",[5] Anderson felt that the show had potential:[7] "Perhaps, with a little more seasoning on the Chicago police department, Jamie Rose might become a star."[7] In his 1991 book The TV Encyclopedia, David Inman called Lady Blue "one of the dumbest shows ever on ABC—and that's saying a lot".[31] Lloyd Grove of The Washington Post criticized the reliance on violence "[that] overpowers, and eventually sours, what could have been an agreeably fast-paced show". In response to the pilot, Grove also panned its writer Robert Vincent O'Neil for copying ideas from Clint Eastwood films and the 1971 film The French Connection.[12] Despite negative reviews, Reesman reported that teenage and young adult males responded positively to Mahoney's attitude and appearance.[1]
In his 2013 book June Cleaver Was a Feminist!: Reconsidering the Female Characters of Early Television, television studies author Cary O'Dell called Lady Blue an "interesting experiment" in imagining the "hardcore cop genre with a female lead". According to O'Dell, criticism of Mahoney and the series' ultimate cancellation were the results of sexism: "Was such rebellion, contempt for authority, and brutal tactics considered too 'unfeminine'?" The author felt that Lady Blue was ahead of its time, contrasting Mahoney's negative reception with the positive reaction to the titular protagonists of the 1991 film Thelma & Louise, who have developed a legacy as "newfangled feminist icons".[24]
Oligarchy (from Ancient Greek ὀλιγαρχία (oligarkhía) 'rule by few'; from ὀλίγος (olígos) 'few' and ἄρχω (árkhō) 'to rule, command')[1][2][3] is a form of government in which power rests with a small number of people. These people may or may not be distinguished by one or several characteristics, such as nobility, fame, wealth, education, or corporate, religious, political, or military control.
Throughout history, power structures considered to be oligarchies have often been viewed as coercive, relying on public obedience or oppression to exist. Aristotle pioneered the use of the term as meaning rule by the rich, contrasting it with aristocracy, arguing that oligarchy was a perversion of aristocracy.[4]
Types
Minority rule
Main article: Minoritarianism
The consolidation of power by a dominant religious or ethnic minority can be considered a form of oligarchy.[5] Examples include South Africa during apartheid, Liberia under Americo-Liberians, the Sultanate of Zanzibar,[citation needed] and Rhodesia. In these cases, oligarchic rule was often tied to the legacy of colonialism.[5]
In the early 20th century, Robert Michels expanded on this idea in his Iron Law of Oligarchy He argued that even democracies, like all large organizations, tend to become oligarchic due to the necessity of dividing labor, which ultimately results in a ruling class focused on maintaining its power.
Putative oligarchies
Main article: Business oligarch
Business groups may be considered oligarchies if they meet the following criteria:
They are the largest private owners in the country.
They possess sufficient political power to influence their own interests.
The owners control multiple businesses, coordinating activities across sectors.[6]
Intellectual oligarchies
George Bernard Shaw coined the concept of an intellectual oligarchy in his play Major Barbara (1907). In the play, Shaw criticizes the control of society by intellectual elites and expresses a desire for the empowerment of the common people:[7]
I now want to give the common man weapons against the intellectual man. I love the common people. I want to arm them against the lawyer, the doctor, the priest, the literary man, the professor, the artist, and the politician, who, once in authority, is the most dangerous, disastrous, and tyrannical of all the fools, rascals, and impostors. I want a democratic power strong enough to force the intellectual oligarchy to use its genius for the general good or else perish.
History
[icon]
This section needs expansion. You can help by adding to it. (January 2024)
By country
Jeffrey A. Winters and Benjamin I. Page have described Colombia, Indonesia, Russia, Singapore and the United States as oligarchies.[8]
The Philippines
Main article: Monopolies in the Philippines (1965–1986)
During the presidency of Ferdinand Marcos from 1965 to 1986, several monopolies arose in the Philippines, primarily linked to the Marcos family and their close associates. Analysts have described this period, and even subsequent decades, as an era of oligarchy in the Philippines.[9][10][11][12]
President Rodrigo Duterte, elected in 2016, promised to dismantle the oligarchy during his presidency.[13][12] However, corporate oligarchy persisted throughout his tenure. While Duterte criticized prominent tycoons such as the Ayalas and Manny Pangilinan, corporate figures allied with Duterte, including Dennis Uy of Udenna Corporation, benefitted during his administration.[14]
Russia
Main article: Russian oligarchs
Since the dissolution of the Soviet Union in 1991 and the subsequent privatization of state-owned assets, a class of Russian oligarchs emerged. These oligarchs gained control of significant portions of the economy, especially in the energy, metals, and natural resources sectors.[15] Many of these individuals maintained close ties with government officials, particularly the president, leading some to characterize modern Russia as an oligarchy intertwined with the state.[16]
Iran
Main articles: Khomeinism and Velayat-e-faqih
The Islamic Republic of Iran, established after the 1979 Iranian Revolution, is sometimes described as a clerical oligarchy. Its ruling system, known as Velayat-e-Faqih (Governance of the Jurists), places power in the hands of a small group of high-ranking Shia clerics, led by the Supreme Leader. This group holds significant influence over the country's legislative, military, and economic affairs, and critics argue that this system concentrates power in a religious elite, marginalizing other voices within society.[17][18]
Ukraine
Main article: Ukrainian oligarchs
Since Ukraine's independence in 1991, a powerful class of business elites, known as Ukrainian oligarchs, has played a significant role in the country's politics and economy. These oligarchs gained control of state assets during the rapid privatization that followed the collapse of the Soviet Union.[6] By 2021, Ukraine passed a law aimed at curbing oligarchic influence on politics and the economy.[19]
United States
Further information: Income inequality in the United States § Democracy and society, and Politics of the United States § Oligarchy
The Bosses of the Senate, corporate interests as giant money bags looming over senators
Several commentators and scholars have suggested that the United States demonstrates characteristics of an oligarchy, particularly in relation to the concentration of wealth and political influence among a small elite,[20][21][22][23][24] as exemplified by the list of top (political party) donors.[25][26][27]
Economist Simon Johnson argued that the rise of an American financial oligarchy became particularly prominent following the 2008 financial crisis.[28] This financial elite has been described as wielding significant power over both the economy and political decisions. Former President Jimmy Carter in 2015 characterized the United States as an "oligarchy with unlimited political bribery" following the 2010 Citizens United v. FEC Supreme Court decision, which removed limits on donations to political campaigns.[29]
In 2014, a study by political scientists Martin Gilens of Princeton University and Benjamin Page of Northwestern University argued that the United States' political system does not primarily reflect the preferences of its average citizens. Their analysis of policy outcomes between 1981 and 2002 suggested that wealthy individuals and business groups held substantial influence over political decisions, often sidelining the majority of Americans.[30] While the United States maintains democratic features such as regular elections, freedom of speech, and widespread suffrage, the study noted that policy decisions are disproportionately influenced by economic elites.[31] However, the study received criticism from other scholars, who argued that the influence of average citizens should not be discounted and that the conclusions about oligarchic tendencies were overstated.[32] Gilens and Page defended their research, reiterating that while they do not label the United States an outright oligarchy, they found substantial evidence of economic elites dominating certain areas of policy-making.[33]
In his presidential farewell address on January 15, 2025, outgoing U.S. President Joe Biden warned that an oligarchy was taking shape in America which threatened democracy, basic rights, and freedom, aided by a tech–industrial complex.[34][35]
Fascism (/ˈfæʃɪzəm/ FASH-iz-əm) is a far-right, authoritarian, and ultranationalist political ideology and movement,[1][2][3] characterized by a dictatorial leader, centralized autocracy, militarism, forcible suppression of opposition, belief in a natural social hierarchy, subordination of individual interests for the perceived good of the nation or race, and strong regimentation of society and the economy.[2][3] Opposed to Marxism, democracy, anarchism, pluralism, free markets, egalitarianism, communism, liberalism, and socialism,[4][5] fascism is at the far right of the traditional left–right spectrum.[6][5][7]
Fascism rose to prominence in early-20th-century Europe.[6][8] The first fascist movements emerged in Italy during World War I, before spreading to other European countries, most notably Germany.[6] Fascism also had adherents outside of Europe.[9] Fascists saw World War I as a revolution that brought massive changes to the nature of war, society, the state, and technology. The advent of total war and the mass mobilization of society erased the distinction between civilians and combatants. A military citizenship arose, in which all citizens were involved with the military in some manner.[10] The war resulted in the rise of a powerful state capable of mobilizing millions of people to serve on the front lines, providing logistics to support them, and having unprecedented authority to intervene in the lives of citizens.[10]
Fascism views forms of violence including political violence, imperialist violence, and war as means to national rejuvenation.[11][12] Fascists often advocate for the establishment of a totalitarian one-party state,[13][14] and for a dirigiste economy (a market economy in which the state plays a strong directive role through economic interventionist policies), with the principal goal of achieving autarky (national economic self-sufficiency).[15][16] Fascism's extreme authoritarianism and nationalism can manifest as a belief in Manifest Destiny or a revival of historical greatness (like Mussolini seeking to restore the Roman Empire). It may also centre around an ingroup-outgroup opposition. In the case of Nazism, this involved racial purity and a master race which blended with a variant of racism and discrimination against a demonized "Other", such as Jews and other groups. Other marginalized groups such as homosexuals, transgender people, ethnic minorities, or immigrants have been targeted. Such bigotry has motivated fascist regimes to commit massacres, forced sterilizations, deportations, and genocides.[17][18] During World War II, the genocidal and imperialist ambitions of the fascist Axis powers resulted in the murder of millions of people.
Since the end of World War II in 1945, fascism has been largely disgraced, and few parties have openly described themselves as fascist; the term is often used pejoratively by political opponents. The descriptions neo-fascist or post-fascist are sometimes applied to contemporary parties with ideologies similar to, or rooted in, 20th-century fascist movements.[6][19] Some opposition groups have adopted the label anti-fascist (often shortened to antifa) to signify their stance.[20]
Etymology
The fasces, a symbol of Ancient Rome, was employed in the modern era by various political movements to denote strength through unity.[21]
The Italian term fascismo is derived from fascio, meaning 'bundle of sticks', ultimately from the Latin word fasces.[3] This was the name given to political organizations in Italy known as fasci, groups similar to guilds or syndicates. According to Italian fascist dictator Benito Mussolini's own account, the Fasces of Revolutionary Action were founded in Italy in 1915.[22] In 1919, Mussolini founded the Italian Fasces of Combat in Milan, which became the National Fascist Party two years later. The fascists came to associate the term with the ancient Roman fasces or fascio littorio,[23] a bundle of rods tied around an axe,[24] an ancient Roman symbol of the authority of the civic magistrate,[25] carried by his lictors.[26] The symbolism of the fasces suggested strength through unity: a single rod is easily broken, while the bundle is difficult to break.[27]
Prior to 1914, the fasces symbol was widely employed by various political movements, often of a left-wing or liberal persuasion. For instance, according to Robert Paxton, "Marianne, symbol of the French Republic, was often portrayed in the nineteenth century carrying the fasces to represent the force of Republican solidarity against her aristocratic and clerical enemies."[21] The symbol often appeared as an architectural motif, for instance on the Sheldonian Theater at Oxford University and on the Lincoln Memorial in Washington, D.C.[21]
Definitions
Main article: Definitions of fascism
Part of a series on
Fascism
Eagle with fasces
Principles
Topics
Politicians
Intellectuals
Literature
Organizations
Media
History
Variants
By continent
Related topics
icon Politics portal
vte
Historian Ian Kershaw once wrote, "Trying to define 'fascism' is like trying to nail jelly to the wall."[28] Each group described as "fascist" has at least some unique elements, and frequently definitions of "fascism" have been criticized as either too broad or too narrow.[29] According to many scholars, fascists—especially when they are in power—have historically attacked communism, socialism, conservatism, and parliamentary liberalism, attracting support primarily from the far-right.[30]
Historian Stanley G. Payne's definition is frequently cited as standard by notable scholars,[31] such as Roger Griffin,[32] Randall Schweller,[33] Bo Rothstein,[34] Federico Finchelstein,[35] and Stephen D. Shenfield,[36][37] His definition of fascism focuses on three concepts:
"Fascist negations" – anti-liberalism, anti-communism, and anti-conservatism.
"Fascist goals" – the creation of a nationalist dictatorship to regulate economic structure and to transform social relations within a modern, self-determined culture, and the expansion of the nation into an empire.
"Fascist style" – a political aesthetic of romantic symbolism, mass mobilization, a positive view of violence, and promotion of masculinity, youth, and charismatic authoritarian leadership.[38]
Umberto Eco lists fourteen "features that are typical of what [he] would like to call 'Ur-Fascism', or 'Eternal Fascism'. These features cannot be organized into a system; many of them contradict each other, and are also typical of other kinds of despotism or fanaticism. But it is enough that one of them be present to allow fascism to coagulate around it."[39] Historian John Lukacs argues that there is no such thing as generic fascism. He claims that Nazism and communism are essentially manifestations of populism, and that states such as Nazi Germany and Fascist Italy are more different from each other than they are similar.[40]
Roundel used on the wings of aircraft of the Italian air force during the Fascist period
In his book How Fascism Works: The Politics of Us and Them (2018), Jason Stanley defined fascism thusly:
[A] cult of the leader who promises national restoration in the face of humiliation brought on by supposed communists, Marxists and minorities and immigrants who are supposedly posing a threat to the character and the history of a nation ... The leader proposes that only he can solve it and all of his political opponents are enemies or traitors.
Stanley says recent global events as of 2020, including the COVID-19 pandemic and the 2020–2022 United States racial unrest, have substantiated his concern about how fascist rhetoric is showing up in politics and policies around the world.[41]
Roger Griffin describes fascism as "a genus of political ideology whose mythic core in its various permutations is a palingenetic form of populist ultranationalism."[42] Without palingenetic ultranationalism, there is no "genuine fascism" according to Griffin.[43] Griffin further describes fascism as having three core components: "(i) the rebirth myth, (ii) populist ultra-nationalism, and (iii) the myth of decadence."[44] In Griffin's view, fascism is "a genuinely revolutionary, trans-class form of anti-liberal, and in the last analysis, anti-conservative nationalism" built on a complex range of theoretical and cultural influences. He distinguishes an inter-war period in which it manifested itself in elite-led but populist "armed party" politics opposing socialism and liberalism, and promising radical politics to rescue the nation from decadence.[45]
Kershaw argues that the difference between fascism and other forms of right-wing authoritarianism in the interwar period is that the latter generally aimed "to conserve the existing social order", whereas fascism was "revolutionary", seeking to change society and obtain "total commitment" from the population.[46] In Against the Fascist Creep, Alexander Reid Ross writes regarding Griffin's view: "Following the Cold War and shifts in fascist organizing techniques, a number of scholars have moved toward the minimalist 'new consensus' refined by Roger Griffin: 'the mythic core' of fascism is 'a populist form of palingenetic ultranationalism.' That means that fascism is an ideology that draws on old, ancient, and even arcane myths of racial, cultural, ethnic, and national origins to develop a plan for the 'new man.'"[47] Griffin himself explored this 'mythic' or 'eliminable' core of fascism with his concept of post-fascism to explore the continuation of Nazism in the modern era.[48] Additionally, other historians have applied this minimalist core to explore proto-fascist movements.[49][50]
Cas Mudde and Cristóbal Rovira Kaltwasser argue that although fascism "flirted with populism ... in an attempt to generate mass support", it is better seen as an elitist ideology. They cite in particular its exaltation of the Leader, the race, and the state, rather than the people. They see populism as a "thin-centered ideology" with a "restricted morphology" that necessarily becomes attached to "thick-centered" ideologies such as fascism, liberalism, or socialism. Thus populism can be found as an aspect of many specific ideologies, without necessarily being a defining characteristic of those ideologies. They refer to the combination of populism, authoritarianism and ultranationalism as "a marriage of convenience".[51]
Robert Paxton says:
[Fascism is] a form of political behavior marked by obsessive preoccupation with community decline, humiliation, or victimhood and by compensatory cults of unity, energy, and purity, in which a mass-based party of committed nationalist militants, working in uneasy but effective collaboration with traditional elites, abandons democratic liberties and pursues with redemptive violence and without ethical or legal restraints goals of internal cleansing and external expansion.[52]
Roger Eatwell defines fascism as "an ideology that strives to forge social rebirth based on a holistic-national radical Third Way",[53] while Walter Laqueur sees the core tenets of fascism as "self-evident: nationalism; social Darwinism; racialism, the need for leadership, a new aristocracy, and obedience; and the negation of the ideals of the Enlightenment and the French Revolution."[54]
Historian Emilio Gentile has defined fascism thusly:
[A] modern political phenomenon, revolutionary, anti-liberal, and anti-Marxist, organized in a militia party with a totalitarian conception of politics and the state, an activist and anti-theoretical ideology, with a mythical, virilistic and anti-hedonistic foundation, sacralized as a secular religion, which affirms the absolute primacy of the nation, understood as an ethnically homogeneous organic community, hierarchically organized in a corporate state, with a bellicose vocation to the politics of greatness, power, and conquest aimed at creating a new order and a new civilization.[55]
Historian and cultural critic Ruth Ben-Ghiat has described fascism as "the original phase of authoritarianism, along with early communism, when a population has undergone huge dislocations or they perceive that there's been changes in society that are very rapid, too rapid for their taste".[56]
Racism was a key feature of German fascism, for which the Holocaust was a high priority. According to The Historiography of Genocide, "In dealing with the Holocaust, it is the consensus of historians that Nazi Germany targeted Jews as a race, not as a religious group."[57] Several historians, such as Umberto Eco,[39] Kevin Passmore,[58] and Moyra Grant,[59] stress racism as a characteristic component of German fascism. Historian Robert Soucy stated, "Hitler envisioned the ideal German society as a Volksgemeinschaft, a racially unified and hierarchically organized body in which the interests of individuals would be strictly subordinate to those of the nation, or Volk."[60] Kershaw noted that common factors of fascism included "the 'cleansing' of all those deemed not to belong—foreigners, ethnic minorities, 'undesirables'" and belief in its own nation's superiority, even if it was not biological racism like in Nazism.[46] Fascist philosophies vary by application, but remain distinct by one theoretical commonality: all traditionally fall into the far-right sector of any political spectrum, catalyzed by afflicted class identities over conventional social inequities.[6]
According to the Council on Foreign Relations, many experts see fascism as a mass political movement centered around extreme nationalism, militarism, and the placement of national interests above those of the individual. Fascist regimes often advocate for the overthrow of institutions that they view as "liberal decay" while simultaneously promoting traditional values. They believe in the supremacy of certain peoples and use it to justify the persecution of other groups. Fascist leaders often maintain a cult of personality and seek to generate enthusiasm for the regime by rallying massive crowds. This contrasts with authoritarian governments, which also centralize power and suppress dissent, but want their subjects to remain passive and demobilized.[61]
Position on the political spectrum
Pro-government demonstration in Salamanca, Francoist Spain, in 1937. Francisco Franco was later labeled by some commentators the "last surviving fascist dictator".[62]
Scholars place fascism on the far right of the political spectrum.[6][5][7] Such scholarship focuses on its social conservatism and its authoritarian means of opposing egalitarianism.[63] Roderick Stackelberg places fascism—including Nazism, which he says is "a radical variant of fascism"—on the political right by explaining: "The more a person deems absolute equality among all people to be a desirable condition, the further left he or she will be on the ideological spectrum. The more a person considers inequality to be unavoidable or even desirable, the further to the right he or she will be."[64]
Fascism's origins are complex and include many seemingly contradictory viewpoints, ultimately centered on a mythos of national rebirth from decadence.[65] Fascism was founded during World War I by Italian national syndicalists who drew upon both left-wing organizational tactics and right-wing political views.[66] Italian fascism gravitated to the right in the early 1920s.[67] A major element of fascist ideology that has been deemed to be far right is its stated goal to promote the right of a supposedly superior people to dominate, while purging society of supposedly inferior elements.[68]
Mussolini and Giovanni Gentile described their ideology as right-wing in the political essay The Doctrine of Fascism (1932), stating: "We are free to believe that this is the century of authority, a century tending to the 'right,' a fascist century."[69] Mussolini stated that fascism's position on the political spectrum was not a serious issue for fascists: "[F]ascism, sitting on the right, could also have sat on the mountain of the center. ... These words in any case do not have a fixed and unchanged meaning: they do have a variable subject to location, time and spirit. We don't give a damn about these empty terminologies and we despise those who are terrorized by these words."[70]
Major Italian groups politically on the right, especially rich landowners and big business, feared an uprising by groups on the left, such as sharecroppers and labour unions.[71] They welcomed fascism and supported its violent suppression of opponents on the left.[72] The accommodation of the political right into the Italian Fascist movement in the early 1920s created internal factions within the movement. The "fascist left" included Michele Bianchi, Giuseppe Bottai, Angelo Oliviero Olivetti, Sergio Panunzio, and Edmondo Rossoni, who were committed to advancing national syndicalism as a replacement for parliamentary liberalism in order to modernize the economy and advance the interests of workers and the common people.[73] The "fascist right" included members of the paramilitary Blackshirts and former members of the Italian Nationalist Association (ANI).[73] The Blackshirts wanted to establish fascism as a complete dictatorship, while the former ANI members, including Alfredo Rocco, sought to institute an authoritarian corporatist state to replace the liberal state in Italy while retaining the existing elites.[73] Upon accommodating the political right, there arose a group of monarchist fascists who sought to use fascism to create an absolute monarchy under King Victor Emmanuel III of Italy.[73]
A number of post-World War II fascist movements described themselves as a Third Position outside the traditional political spectrum. Falange Española de las JONS leader José Antonio Primo de Rivera said: "[B]asically the Right stands for the maintenance of an economic structure, albeit an unjust one, while the Left stands for the attempt to subvert that economic structure, even though the subversion thereof would entail the destruction of much that was worthwhile."[74]
Fascist as a pejorative
Main article: Fascist (insult)
The term fascist has been used as a pejorative,[75] regarding varying movements across the far right of the political spectrum. George Orwell noted in 1944 that the term had been used to denigrate diverse positions "in internal politics". Orwell said that while fascism is "a political and economic system" that was inconvenient to define, "as used, the word 'Fascism' is almost entirely meaningless. ... almost any English person would accept 'bully' as a synonym for 'Fascist'",[76] and in 1946 wrote that "'Fascism' has now no meaning except in so far as it signifies something not desirable."[77] Richard Griffiths of the University of Wales wrote in 2000 that "fascism" is the "most misused, and over-used word, of our times".[78]: 1 Fascist is sometimes applied to post-World War II organizations and ways of thinking that academics more commonly term neo-fascist.[79]
Despite fascist movements' history of anti-communism, Communist states have sometimes been referred to as fascist, typically as an insult. It has been applied to Marxist–Leninist regimes in Cuba under Fidel Castro and Vietnam under Ho Chi Minh.[80] Chinese Marxists used the term to denounce the Soviet Union during the Sino-Soviet split, and the Soviets used the term to denounce Chinese Marxists,[81] in addition to social democracy, coining a new term in social fascism. In the United States, Herbert Matthews of The New York Times asked in 1946: "Should we now place Stalinist Russia in the same category as Hitlerite Germany? Should we say that she is Fascist?"[82] J. Edgar Hoover, longtime FBI director and ardent anti-communist, wrote extensively of red fascism.[83] The Ku Klux Klan in the 1920s was sometimes called fascist. Historian Peter Amann states that, "Undeniably, the Klan had some traits in common with European fascism—chauvinism, racism, a mystique of violence, an affirmation of a certain kind of archaic traditionalism—yet their differences were fundamental ... [the KKK] never envisioned a change of political or economic system."[84]
History
Further information: Fascism and ideology
Background and 19th-century roots
Depiction of a Greek Hoplite warrior; ancient Sparta has been considered an inspiration for fascist and quasi-fascist movements, such as Nazism and quasi-fascist Metaxism[85]
Early influences that shaped the ideology of fascism have been dated back to ancient Greece. The political culture of ancient Greece and specifically the ancient Greek city state of Sparta under Lycurgus, with its emphasis on militarism and racial purity, were admired by the Nazis.[86][87][88] Nazi Führer Adolf Hitler emphasized that Germany should adhere to Hellenic values and culture – particularly that of ancient Sparta.[86][87]
Georges Valois, founder of the first non-Italian fascist party Faisceau,[89] claimed the roots of fascism stemmed from the late 18th century Jacobin movement, seeing in its totalitarian nature a foreshadowing of the fascist state.[90] Historian George Mosse similarly analyzed fascism as an inheritor of the mass ideology and civil religion of the French Revolution, as well as a result of the brutalization of societies in 1914–1918.[90]
Historians such as Irene Collins and Howard C Payne see Napoleon III, who ran a 'police state' and suppressed the media, as a forerunner of fascism.[91] According to David Thomson,[92] the Italian Risorgimento of 1871 led to the 'nemesis of fascism'. William L Shirer[93] sees a continuity from the views of Fichte and Hegel, through Bismarck, to Hitler; Robert Gerwarth speaks of a 'direct line' from Bismarck to Hitler.[94] Julian Dierkes sees fascism as a 'particularly violent form of imperialism'.[95]
Marcus Garvey, founder and leader of the Universal Negro Improvement Association, insisted that he and his organisation "were the first fascists".[96] In 1938, C. L. R. James wrote "all the things that Hitler was to do so well later, Marcus Garvey was doing in 1920 and 1921".[97]
Fin de siècle era and lead up to World War I (1880–1914)
See also: National syndicalism
The historian Zeev Sternhell has traced the ideological roots of fascism back to the 1880s and in particular to the fin de siècle theme of that time.[98] The theme was based on a revolt against materialism, rationalism, positivism, bourgeois society, and democracy.[99] The fin-de-siècle generation supported emotionalism, irrationalism, subjectivism, and vitalism.[100] They regarded civilization as being in crisis, requiring a massive and total solution.[99] Their intellectual school considered the individual as only one part of the larger collectivity, which should not be viewed as a numerical sum of atomized individuals.[99] They condemned the rationalistic, liberal individualism of society and the dissolution of social links in bourgeois society.[99]
The fin-de-siècle outlook was influenced by various intellectual developments, including Darwinian biology, Gesamtkunstwerk, Arthur de Gobineau's racialism, Gustave Le Bon's psychology, and the philosophies of Friedrich Nietzsche, Fyodor Dostoyevsky, and Henri Bergson.[101] Social Darwinism, which gained widespread acceptance, made no distinction between physical and social life, and viewed the human condition as being an unceasing struggle to achieve the survival of the fittest.[101] It challenged positivism's claim of deliberate and rational choice as the determining behaviour of humans, with social Darwinism focusing on heredity, race, and environment.[101] Its emphasis on biogroup identity and the role of organic relations within societies fostered the legitimacy and appeal of nationalism.[102] New theories of social and political psychology also rejected the notion of human behaviour being governed by rational choice and instead claimed that emotion was more influential in political issues than reason.[101] Nietzsche's argument that "God is dead" coincided with his attack on the "herd mentality" of Christianity, democracy, and modern collectivism, his concept of the Übermensch, and his advocacy of the will to power as a primordial instinct, were major influences upon many of the fin-de-siècle generation.[103] Bergson's claim of the existence of an élan vital, or vital instinct, centred upon free choice and rejected the processes of materialism and determinism; this challenged Marxism.[104]
In his work The Ruling Class (1896), Gaetano Mosca developed the theory that claims that in all societies an "organized minority" would dominate and rule over an "disorganized majority",[105] stating that there are only two classes in society, "the governing" (the organized minority) and "the governed" (the disorganized majority).[106] He claims that the organized nature of the organized minority makes it irresistible to any individual of the disorganized majority.[106]
Charles Maurras
Georges Sorel
French nationalist and reactionary monarchist Charles Maurras influenced fascism.[107] Maurras promoted what he called integral nationalism, which called for the organic unity of a nation, and insisted that a powerful monarch was an ideal leader of a nation. Maurras distrusted what he considered the democratic mystification of the popular will that created an impersonal collective subject.[107] He claimed that a powerful monarch was a personified sovereign who could exercise authority to unite a nation's people.[107] Maurras' integral nationalism was idealized by fascists, but modified into a modernized revolutionary form that was devoid of Maurras' monarchism.[107]
French revolutionary syndicalist Georges Sorel promoted the legitimacy of political violence in his work Reflections on Violence (1908) and other works in which he advocated radical syndicalist action to achieve a revolution to overthrow capitalism and the bourgeoisie through a general strike.[108] In Reflections on Violence, Sorel emphasized need for a revolutionary political religion.[109] Also in his work The Illusions of Progress, Sorel denounced democracy as reactionary, saying "nothing is more aristocratic than democracy."[110] By 1909, after the failure of a syndicalist general strike in France, Sorel and his supporters left the radical left and went to the radical right, where they sought to merge militant Catholicism and French patriotism with their views—advocating anti-republican Christian French patriots as ideal revolutionaries.[111] Initially, Sorel had officially been a revisionist of Marxism, but by 1910 announced his abandonment of socialist literature and claimed in 1914, using an aphorism of Benedetto Croce that "socialism is dead" because of the "decomposition of Marxism".[112] Sorel became a supporter of reactionary Maurrassian nationalism beginning in 1909 that influenced his works.[112] Maurras held interest in merging his nationalist ideals with Sorelian syndicalism, known as Sorelianism, as a means to confront democracy.[113] Maurras stated, "A socialism liberated from the democratic and cosmopolitan element fits nationalism well as a well made glove fits a beautiful hand."[114]
The fusion of Maurrassian nationalism and Sorelian syndicalism influenced radical Italian nationalist Enrico Corradini.[115] Corradini spoke of the need for a nationalist-syndicalist movement, led by elitist aristocrats and anti-democrats who shared a revolutionary syndicalist commitment to direct action and a willingness to fight.[115] Corradini spoke of Italy as being a "proletarian nation" that needed to pursue imperialism in order to challenge the "plutocratic" French and British.[116] Corradini's views were part of a wider set of perceptions within the right-wing Italian Nationalist Association (ANI), which claimed that Italy's economic backwardness was caused by corruption in its political class, liberalism, and division caused by "ignoble socialism".[116]
The ANI held ties and influence among conservatives, Catholics, and the business community.[117] Italian national syndicalists held a common set of principles: the rejection of bourgeois values, democracy, liberalism, Marxism, internationalism, and pacifism, and the promotion of heroism, vitalism, and violence.[118] The ANI claimed that liberal democracy was no longer compatible with the modern world, and advocated a strong state and imperialism. They believed that humans are naturally predatory, and that nations are in a constant struggle in which only the strongest would survive.[119]
Filippo Tommaso Marinetti, Italian modernist author of the Futurist Manifesto (1909) and later the co-author of the Fascist Manifesto (1919)
Futurism was both an artistic-cultural movement and initially a political movement in Italy led by Filippo Tommaso Marinetti who founded the Manifesto of Futurism (1908), that championed the causes of modernism, action, and political violence as necessary elements of politics while denouncing liberalism and parliamentary politics. Marinetti rejected conventional democracy based on majority rule and egalitarianism, for a new form of democracy, promoting what he described in his work "The Futurist Conception of Democracy" as the following: "We are therefore able to give the directions to create and to dismantle to numbers, to quantity, to the mass, for with us number, quantity and mass will never be—as they are in Germany and Russia—the number, quantity and mass of mediocre men, incapable and indecisive."[120]
Futurism influenced fascism in its emphasis on recognizing the virile nature of violent action and war as being necessities of modern civilization.[121] Marinetti promoted the need of physical training of young men saying that, in male education, gymnastics should take precedence over books. He advocated segregation of the genders because womanly sensibility must not enter men's education, which he claimed must be "lively, bellicose, muscular and violently dynamic."[122]
World War I and its aftermath (1914–1929)
Benito Mussolini in 1917 as an Italian soldier in World War I.
At the outbreak of World War I in August 1914, the Italian political left became severely split over its position on the war. The Italian Socialist Party (PSI) opposed the war but a number of Italian revolutionary syndicalists supported war against Germany and Austria-Hungary on the grounds that their reactionary regimes had to be defeated to ensure the success of socialism.[123] Angelo Oliviero Olivetti formed a pro-interventionist fascio called the Revolutionary Fasces of International Action in October 1914.[123] Benito Mussolini upon being expelled from his position as chief editor of the PSI's newspaper Avanti! for his anti-German stance, joined the interventionist cause in a separate fascio.[124] The term "fascism" was first used in 1915 by members of Mussolini's movement, the Fasces of Revolutionary Action.[125]
The first meeting of the Fasces of Revolutionary Action was held on 24 January 1915[126] when Mussolini declared that it was necessary for Europe to resolve its national problems—including national borders—of Italy and elsewhere "for the ideals of justice and liberty for which oppressed peoples must acquire the right to belong to those national communities from which they descended."[126] Attempts to hold mass meetings were ineffective and the organization was regularly harassed by government authorities and socialists.[127]
Adolf Hitler as a German soldier in World War I.
Similar political ideas arose in Germany after the outbreak of the war. German sociologist Johann Plenge spoke of the rise of a "National Socialism" in Germany within what he termed the "ideas of 1914" that were a declaration of war against the "ideas of 1789" (the French Revolution).[128] According to Plenge, the "ideas of 1789"—such as the rights of man, democracy, individualism and liberalism—were being rejected in favor of "the ideas of 1914" that included "German values" of duty, discipline, law and order.[128] Plenge believed that racial solidarity (Volksgemeinschaft) would replace class division and that "racial comrades" would unite to create a socialist society in the struggle of "proletarian" Germany against "capitalist" Britain.[128] He believed that the Spirit of 1914 manifested itself in the concept of the People's League of National Socialism.[129] This National Socialism was a form of state socialism that rejected the "idea of boundless freedom" and promoted an economy that would serve the whole of Germany under the leadership of the state.[129] This National Socialism was opposed to capitalism because of the components that were against "the national interest" of Germany but insisted that National Socialism would strive for greater efficiency in the economy.[129] Plenge advocated an authoritarian rational ruling elite to develop National Socialism through a hierarchical technocratic state.[130]
Impact of World War I
Members of Italy's Arditi corps, shown here in 1918 holding daggers, a symbol of their group. They were formed in 1917 as groups of soldiers trained for dangerous missions, characterized by a refusal to surrender and a willingness to fight to the death. Their black uniforms inspired those of the Italian Fascist movement.
Fascists viewed World War I as bringing revolutionary changes in the nature of war, society, the state and technology, as the advent of total war and mass mobilization had broken down the distinction between civilian and combatant, as civilians had become a critical part in economic production for the war effort and thus arose a "military citizenship" in which all citizens were involved to the military in some manner during the war.[10] World War I had resulted in the rise of a powerful state capable of mobilizing millions of people to serve on the front lines or provide economic production and logistics to support those on the front lines, as well as having unprecedented authority to intervene in the lives of citizens.[10] Fascists viewed technological developments of weaponry and the state's total mobilization of its population in the war as symbolizing the beginning of a new era fusing state power with mass politics, technology and particularly the mobilizing myth that they contended had triumphed over the myth of progress and the era of liberalism.[131]
Impact of the October Revolution in Russia
See also: October Revolution
The October Revolution of 1917, in which Bolshevik communists led by Vladimir Lenin seized power in Russia, greatly influenced the development of fascism.[132] In 1917, Mussolini, as leader of the Fasces of Revolutionary Action, praised the October Revolution, but later he became unimpressed with Lenin, regarding him as merely a new version of Tsar Nicholas II.[133] After World War I, fascists commonly campaigned on anti-Marxist agendas.[132]
British historian Cyprian Blamires argues that there are similarities between fascism and Bolshevism, including that they believed in the necessity of a vanguard leadership, showed contempt for bourgeois values, and had totalitarian ambitions.[132] In practice, both have commonly emphasized revolutionary action, proletarian nation theories, one-party states, and party-armies;[132] With the antagonism between anti-interventionist Marxists and pro-interventionist fascists complete by the end of the war, the two sides became irreconcilable. The fascists presented themselves as anti-communists and as especially opposed to the Marxists.[134] In 1919, Mussolini consolidated control over the fascist movement, known as Sansepolcrismo, with the founding of the Italian Fasces of Combat.[72]
Fascist Manifesto and Charter of Carnaro
Territories promised to Italy by the Treaty of London (1915): Trentino-Alto Adige, the Julian March and Dalmatia (tan) and the Snežnik Plateau area (green). However, after World War I, Dalmatia was not assigned to Italy but to Yugoslavia
In 1919, Alceste De Ambris and futurist movement leader Filippo Tommaso Marinetti created "The Manifesto of the Italian Fasces of Combat".[135] The Fascist Manifesto was presented on 6 June 1919 in the fascist newspaper Il Popolo d'Italia and supported the creation of universal suffrage, including women's suffrage (the latter being realized only partly in late 1925, with all opposition parties banned or disbanded);[136] proportional representation on a regional basis; government representation through a corporatist system of "National Councils" of experts, selected from professionals and tradespeople, elected to represent and hold legislative power over their respective areas, including labour, industry, transportation, public health, and communications, among others; and abolition of the Senate of the Kingdom of Italy.[137] The Fascist Manifesto supported the creation of an eight-hour work day for all workers, a minimum wage, worker representation in industrial management, equal confidence in labour unions as in industrial executives and public servants, reorganization of the transportation sector, revision of the draft law on invalidity insurance, reduction of the retirement age from 65 to 55, a strong progressive tax on capital, confiscation of the property of religious institutions and abolishment of bishoprics, and revision of military contracts to allow the government to seize 85% of profits.[138] It also called for the fulfillment of expansionist aims in the Balkans and other parts of the Mediterranean, the creation of a short-service national militia to serve defensive duties, nationalization of the armaments industry, and a foreign policy designed to be peaceful but also competitive.[139]
Residents of Fiume cheer the arrival of Gabriele d'Annunzio and his blackshirt-wearing nationalist raiders, as D'Annunzio and fascist Alceste De Ambris developed the quasi-fascist Italian Regency of Carnaro (a city-state in Fiume) from 1919 to 1920 and whose actions inspired the Italian fascist movement. In September 1919 Fiume had 22,488 (62% of the population) Italians in a total population of 35,839 inhabitants
The next events that influenced the fascists in Italy were the raid of Fiume by Italian nationalist Gabriele d'Annunzio and the founding of the Charter of Carnaro in 1920.[140] D'Annunzio and De Ambris designed the Charter, which advocated national-syndicalist corporatist productionism alongside D'Annunzio's political views.[141] Many fascists saw the Charter of Carnaro as an ideal constitution for a fascist Italy.[142] This behaviour of aggression towards Yugoslavia and South Slavs was pursued by Italian fascists with their persecution of South Slavs—especially Slovenes and Croats.
From populism to conservative accommodations
In 1920, militant strike activity by industrial workers reached its peak in Italy and 1919 and 1920 were known as the "Red Year" (Biennio Rosso).[143] Mussolini and the fascists took advantage of the situation by allying with industrial businesses and attacking workers and peasants in the name of preserving order and internal peace in Italy.[144]
Fascists identified their primary opponents as the majority of socialists on the left who had opposed intervention in World War I.[142] The fascists and the Italian political right held common ground: both held Marxism in contempt, discounted class consciousness and believed in the rule of elites.[145] The fascists assisted the anti-socialist campaign by allying with the other parties and the conservative right in a mutual effort to destroy the Italian Socialist Party and labour organizations committed to class identity above national identity.[145]
Fascism sought to accommodate Italian conservatives by making major alterations to its political agenda—abandoning its previous populism, republicanism and anticlericalism, adopting policies in support of free enterprise and accepting the Catholic Church and the monarchy as institutions in Italy.[146] To appeal to Italian conservatives, fascism adopted policies such as promoting family values, including policies designed to reduce the number of women in the workforce—limiting the woman's role to that of a mother. The fascists banned literature on birth control and increased penalties for abortion in 1926, declaring both crimes against the state.[147]
Although fascism adopted a number of anti-modern positions designed to appeal to people upset with the new trends in sexuality and women's rights—especially those with a reactionary point of view—the fascists sought to maintain fascism's revolutionary character, with Angelo Oliviero Olivetti saying: "Fascism would like to be conservative, but it will [be] by being revolutionary."[148] The Fascists supported revolutionary action and committed to secure law and order to appeal to both conservatives and syndicalists.[149]
Prior to fascism's accommodations to the political right, fascism was a small, urban, northern Italian movement that had about a thousand members.[150] After Fascism's accommodation of the political right, the fascist movement's membership soared to approximately 250,000 by 1921.[151] A 2020 article by Daron Acemoğlu, Giuseppe De Feo, Giacomo De Luca, and Gianluca Russo in the Center for Economic and Policy Research, exploring the link between the threat of socialism and Mussolini's rise to power, found "a strong association between the Red Scare in Italy and the subsequent local support for the Fascist Party in the early 1920s." According to the authors, it was local elites and large landowners who played an important role in boosting Fascist Party activity and support, which did not come from socialists' core supporters but from centre-right voters, as they viewed traditional centre-right parties as ineffective in stopping socialism and turned to the Fascists. In 2003, historian Adrian Lyttelton wrote: "The expansion of Fascism in the rural areas was stimulated and directed by the reaction of the farmers and landowners against the peasant leagues of both Socialists and Catholics."[152]
Fascist violence
Beginning in 1922, fascist paramilitaries escalated their strategy from one of attacking socialist offices and the homes of socialist leadership figures, to one of violent occupation of cities. The fascists met little serious resistance from authorities and proceeded to take over several northern Italian cities.[153] The fascists attacked the headquarters of socialist and Catholic labour unions in Cremona and imposed forced Italianization upon the German-speaking population of Bolzano.[153][154] After seizing these cities, the fascists made plans to take Rome.[153]
Benito Mussolini with three of the four quadrumvirs during the March on Rome (from left to right: unknown, de Bono, Mussolini, Balbo and de Vecchi)
On 24 October 1922, the Fascist Party held its annual congress in Naples, where Mussolini ordered Blackshirts to take control of public buildings and trains and to converge on three points around Rome.[153] The Fascists managed to seize control of several post offices and trains in northern Italy while the Italian government, led by a left-wing coalition, was internally divided and unable to respond to the Fascist advances.[155] King Victor Emmanuel III of Italy perceived the risk of bloodshed in Rome in response to attempting to disperse the Fascists to be too high.[156] Victor Emmanuel III decided to appoint Mussolini as Prime Minister of Italy and Mussolini arrived in Rome on 30 October to accept the appointment.[156] Fascist propaganda aggrandized this event, known as "March on Rome", as a "seizure" of power because of Fascists' heroic exploits.[153]
Fascist Italy
Historian Stanley G. Payne says:
[Fascism in Italy was a] primarily political dictatorship. ... The Fascist Party itself had become almost completely bureaucratized and subservient to, not dominant over, the state itself. Big business, industry, and finance retained extensive autonomy, particularly in the early years. The armed forces also enjoyed considerable autonomy. ... The Fascist militia was placed under military control. ... The judicial system was left largely intact and relatively autonomous as well. The police continued to be directed by state officials and were not taken over by party leaders ... nor was a major new police elite created. ... There was never any question of bringing the Church under overall subservience. ... Sizable sectors of Italian cultural life retained extensive autonomy, and no major state propaganda-and-culture ministry existed. ... The Mussolini regime was neither especially sanguinary nor particularly repressive.[157]
Mussolini in power
Italian ethnic regions claimed in the 1930s. Savoy and Corfu were later claimed.
Nice, Ticino and Dalmatia
Malta
Corsica
Upon being appointed Prime Minister of Italy, Mussolini had to form a coalition government because the fascists did not have control over the Italian parliament.[158] Mussolini's coalition government initially pursued economically liberal policies under the direction of liberal finance minister Alberto De Stefani, a member of the Center Party, including balancing the budget through deep cuts to the civil service.[158] Initially, little drastic change in government policy had occurred and repressive police actions were limited.[158]
The fascists began their attempt to entrench fascism in Italy with the Acerbo Law, which guaranteed a plurality of the seats in parliament to any party or coalition list in an election that received 25% or more of the vote.[159] Through considerable fascist violence and intimidation, the list won a majority of the vote, allowing many seats to go to the fascists.[159] In the aftermath of the election, a crisis and political scandal erupted after Socialist Party deputy Giacomo Matteotti was kidnapped and murdered by a Fascist.[159] The liberals and the leftist minority in parliament walked out in protest in what became known as the Aventine Secession.[160] On 3 January 1925, Mussolini addressed the Fascist-dominated Italian parliament and declared that he was personally responsible for what happened, but insisted that he had done nothing wrong. Mussolini proclaimed himself dictator of Italy, assuming full responsibility over the government and announcing the dismissal of parliament.[160] From 1925 to 1929, fascism steadily became entrenched in power: opposition deputies were denied access to parliament, censorship was introduced and a December 1925 decree made Mussolini solely responsible to the King.[161]
Catholic Church
The signing of the Lateran Treaty, Mussolini shown on the right side of the photograph.
In 1929, the fascist regime briefly gained what was in effect a blessing of the Catholic Church after the regime signed a concordat with the Church, known as the Lateran Treaty, which gave the papacy state sovereignty and financial compensation for the seizure of Church lands by the liberal state in the 19th century, but within two years the Church had renounced fascism in the Encyclical Non Abbiamo Bisogno as a "pagan idolatry of the state" which teaches "hatred, violence and irreverence".[162] Not long after signing the agreement, by Mussolini's own confession, the Church had threatened to have him "excommunicated", in part because of his intractable nature, but also because he had "confiscated more issues of Catholic newspapers in the next three months than in the previous seven years."[163] By the late 1930s, Mussolini became more vocal in his anti-clerical rhetoric, repeatedly denouncing the Catholic Church and discussing ways to depose the pope. He took the position that the "papacy was a malignant tumor in the body of Italy and must 'be rooted out once and for all,' because there was no room in Rome for both the Pope and himself."[164] In her 1974 book, Mussolini's widow Rachele stated that her husband had always been an atheist until near the end of his life, writing that her husband was "basically irreligious until the later years of his life."[165]
The Nazis in Germany employed similar anti-clerical policies. The Gestapo confiscated hundreds of monasteries in Austria and Germany, evicted clergymen and laymen alike and often replaced crosses with swastikas.[166] Referring to the swastika as "the Devil's Cross", church leaders found their youth organizations banned, their meetings limited and various Catholic periodicals censored or banned. Government officials eventually found it necessary to place "Nazis into editorial positions in the Catholic press."[167] Up to 2,720 clerics, mostly Catholics, were arrested by the Gestapo and imprisoned inside of Germany's Dachau concentration camp, resulting in over 1,000 deaths.[168]
Corporatist economic system
The fascist regime created a corporatist economic system in 1925 with creation of the Palazzo Vidoni Pact, in which the Italian employers' association Confindustria and fascist trade unions agreed to recognize each other as the sole representatives of Italy's employers and employees, excluding non-fascist trade unions.[169] The Fascist regime first created a Ministry of Corporations that organized the Italian economy into 22 sectoral corporations, banned workers' strikes and lock-outs and in 1927 created the Charter of Labour, which established workers' rights and duties and created labour tribunals to arbitrate employer-employee disputes.[169] In practice, the sectoral corporations exercised little independence and were largely controlled by the regime, and the employee organizations were rarely led by employees themselves, but instead by appointed Fascist party members.[169]
Aggressive foreign policy
In the 1920s, Fascist Italy pursued an aggressive foreign policy that included ambitions to expand Italian territory.[170] In response to revolt in the Italian colony of Libya, Fascist Italy abandoned previous liberal-era colonial policy of cooperation with local leaders. Instead, claiming that Italians were a superior race to African races and thereby had the right to colonize the "inferior" Africans, it sought to settle 10 to 15 million Italians in Libya.[171] This resulted in an aggressive military campaign known as the Pacification of Libya against natives in Libya, including mass killings, the use of concentration camps and the forced starvation of thousands of people.[171] Italian authorities committed ethnic cleansing by forcibly expelling 100,000 Bedouin Cyrenaicans, half the population of Cyrenaica in Libya, from their settlements that was slated to be given to Italian settlers.[172]
Nazi adoption of the Italian model
Nazis in Munich during the Beer Hall Putsch
The March on Rome brought fascism international attention. One early admirer of the Italian fascists was Adolf Hitler, who less than a month after the March had begun to model himself and the Nazi Party upon Mussolini and the Fascists.[173] The Nazis, led by Hitler and the German war hero Erich Ludendorff, attempted a "March on Berlin" modeled upon the March on Rome, which resulted in the failed Beer Hall Putsch in Munich in November 1923.[174]
International impact of the Great Depression and buildup to World War II
The conditions of economic hardship caused by the Great Depression brought about an international surge of social unrest.[175] Fascist propaganda blamed the problems of the long depression of the 1930s on minorities and scapegoats: "Judeo-Masonic-bolshevik" conspiracies, left-wing internationalism and the presence of immigrants.[176]
In Germany, it contributed to the rise of the Nazi Party, which resulted in the demise of the Weimar Republic and the establishment of the fascist regime, Nazi Germany, under the leadership of Adolf Hitler. With the rise of Hitler and the Nazis to power in 1933, liberal democracy was dissolved in Germany and the Nazis mobilized the country for war, with expansionist territorial aims against several countries. In the 1930s, the Nazis implemented racial laws that deliberately discriminated against, disenfranchised and persecuted Jews and other racial and minority groups.
Jacques Doriot, leader of the French Popular Party speaking at the party's first congress in 1936.
Fascist movements grew in strength elsewhere in Europe. Hungarian fascist Gyula Gömbös rose to power as Prime Minister of Hungary in 1932 and attempted to entrench his Party of National Unity throughout the country. He created an eight-hour work day and a forty-eight-hour work week in industry; sought to entrench a corporatist economy; and pursued irredentist claims on Hungary's neighbors.[177] The fascist Iron Guard movement in Romania soared in political support after 1933, gaining representation in the Romanian government, and an Iron Guard member assassinated Romanian prime minister Ion Duca.[178] The Iron Guard was the only fascist movement outside Germany and Italy to come to power without foreign assistance.[179][180] During the 6 February 1934 crisis, France faced the greatest domestic political turmoil since the Dreyfus Affair when the fascist Francist Movement and multiple far-right movements rioted en masse in Paris against the French government resulting in major political violence.[181] A variety of para-fascist governments that borrowed elements from fascism were formed during the Great Depression, including those of Greece, Lithuania, Poland and Yugoslavia.[182] In the Netherlands, the National Socialist Movement in the Netherlands was at its height in the 1930s due to the Great Depression, especially in 1935 when it won almost eight percent of votes, until the year 1937.[12]
Integralists marching in Brazil
Luis A. Flores, Prime Minister of Peru in 1932, shown saluting in the party uniform of the Revolutionary Union of Peru that he led as its Supreme Chief from 1933-56.
In the Americas, the Brazilian Integralists led by Plínio Salgado claimed as many as 200,000 members, although following coup attempts it faced a crackdown from the Estado Novo of Getúlio Vargas in 1937.[183] In Peru, the Revolutionary Union was a fascist political party which was in power 1931 to 1933. In the 1930s, the National Socialist Movement of Chile gained seats in Chile's parliament and attempted a coup d'état that resulted in the Seguro Obrero massacre of 1938.[184]
During the Great Depression, Mussolini promoted active state intervention in the economy. He denounced the contemporary "supercapitalism" that he claimed began in 1914 as a failure because of its alleged decadence, its support for unlimited consumerism, and its intention to create the "standardization of humankind."[185] Fascist Italy created the Institute for Industrial Reconstruction (IRI), a giant state-owned firm and holding company that provided state funding to failing private enterprises.[186] The IRI was made a permanent institution in Fascist Italy in 1937, pursued fascist policies to create national autarky and had the power to take over private firms to maximize war production.[186] While Hitler's regime only nationalized 500 companies in key industries by the early 1940s,[187] Mussolini declared in 1934, "[t]hree-fourths of Italian economy, industrial and agricultural, is in the hands of the state."[188]
Due to the worldwide depression, Mussolini's government was able to take over most of Italy's largest failing banks, who held controlling interest in many Italian businesses. The IRI reported in early 1934 that they held assets of "48.5 percent of the share capital of Italy", which later included the capital of the banks themselves.[189] Political historian Martin Blinkhorn estimated Italy's scope of state intervention and ownership "greatly surpassed that in Nazi Germany, giving Italy a public sector second only to that of Stalin's Russia."[190] In the late 1930s, Italy enacted manufacturing cartels, tariff barriers, currency restrictions and massive regulation of the economy to attempt to balance payments.[191] Italy's policy of autarky failed to achieve effective economic autonomy.[191] Nazi Germany similarly pursued an economic agenda with the aims of autarky and rearmament and imposed protectionist policies, including forcing the German steel industry to use lower-quality German iron ore rather than superior-quality imported iron.[192]
World War II (1939–1945)
The Greater Germanic Reich, to be realised with the policies of Lebensraum, had boundaries derived from the plans of the Generalplan Ost, the state administration, and the Schutzstaffel (SS).[193]
In Fascist Italy and Nazi Germany, both Mussolini and Hitler pursued territorial expansionist and interventionist foreign policy agendas from the 1930s through the 1940s culminating in World War II. From 1935 to 1939, Germany and Italy escalated their demands for territorial claims and greater influence in world affairs. Italy invaded Ethiopia in 1935 resulting in its condemnation by the League of Nations and its widespread diplomatic isolation. In 1936, Germany remilitarized the industrial Rhineland, a region that had been ordered demilitarized by the Treaty of Versailles. In 1938, Germany annexed Austria and Italy assisted Germany in resolving the diplomatic crisis between Germany versus Britain and France over claims on Czechoslovakia by arranging the Munich Agreement that gave Germany the Sudetenland and was perceived at the time to have averted a European war. These hopes faded when Czechoslovakia was dissolved by the proclamation of the German client state of Slovakia, followed by the next day of the occupation of the remaining Czech Lands and the proclamation of the German Protectorate of Bohemia and Moravia. At the same time from 1938 to 1939, Italy was demanding territorial and colonial concessions from France and Britain.[194] In 1939, Germany prepared for war with Poland, but attempted to gain territorial concessions from Poland through diplomatic means.[195] The Polish government did not trust Hitler's promises and refused to accept Germany's demands.[195]
Map of Great Italy according to the 1940 fascist project in case Italy had won World War II (the orange line delimits metropolitan Italy, the green line the borders of the enlarged Italian Empire)
The invasion of Poland by Germany was deemed unacceptable by Britain, France and their allies, leading to their mutual declaration of war against Germany and the start of World War II. In 1940, Mussolini led Italy into World War II on the side of the Axis. During World War II, the Axis Powers in Europe led by Nazi Germany participated in the extermination of millions of Poles, Jews, Gypsies and others in the genocide known as the Holocaust. In 1943, after Italy faced multiple military failures, the complete reliance and subordination of Italy to Germany, the Allied invasion of Italy and the corresponding international humiliation, Mussolini was removed as head of government and arrested on the order of King Victor Emmanuel III, who proceeded to dismantle the Fascist state and declared Italy's switching of allegiance to the Allied side. Mussolini was rescued from arrest by German forces and led the German client state, the Italian Social Republic from 1943 to 1945. Nazi Germany faced multiple losses and steady Soviet and Western Allied offensives from 1943 to 1945.[196][197]
On 28 April 1945, Mussolini was captured and executed by Italian communist partisans. On 30 April 1945, Hitler committed suicide. Shortly afterwards, Germany surrendered and the Nazi regime was systematically dismantled by the occupying Allied powers. An International Military Tribunal was subsequently convened in Nuremberg. Beginning in November 1945 and lasting through 1949, numerous Nazi political, military and economic leaders were tried and convicted of war crimes, with many of the worst offenders being sentenced to death and executed.[198][199]
Post-World War II (1945–2008)
Main article: Neo-fascism
Juan Perón, President of Argentina from 1946 to 1955 and 1973 to 1974, admired Italian Fascism and modelled his economic policies on those pursued by Fascist Italy.
The victory of the Allies over the Axis powers in World War II led to the collapse of many fascist regimes in Europe. The Nuremberg Trials convicted several Nazi leaders of crimes against humanity involving the Holocaust. However, there remained several movements and governments that were ideologically related to fascism.[200][201]
Francisco Franco's Falangist one-party state in Spain was officially neutral during World War II, although Franco's rise to power had been directly assisted by the militaries of Fascist Italy and Nazi Germany during the Spanish Civil War. The first years were characterized by a repression against the anti-fascist ideologies, deep censorship and the suppression of democratic institutions (elected Parliament, Spanish Constitution of 1931, Regional Statutes of Autonomy). After World War II and a period of international isolation, Franco's regime normalized relations with the Western powers during the Cold War, until Franco's death in 1975 and the transformation of Spain into a liberal democracy.[202]
Historian Robert Paxton observes that one of the main problems in defining fascism is that it was widely mimicked. Paxton says: "In fascism's heyday, in the 1930s, many regimes that were not functionally fascist borrowed elements of fascist decor in order to lend themselves an aura of force, vitality, and mass mobilization." He goes on to observe that Salazar "crushed Portuguese fascism after he had copied some of its techniques of popular mobilization."[203] Paxton says: "Where Franco subjected Spain's fascist party to his personal control, Salazar abolished outright in July 1934 the nearest thing Portugal had to an authentic fascist movement, Rolão Preto's blue-shirted National Syndicalists. ... Salazar preferred to control his population through such 'organic' institutions traditionally powerful in Portugal as the Church. Salazar's regime was not only non-fascist, but 'voluntarily non-totalitarian,' preferring to let those of its citizens who kept out of politics 'live by habit.'"[204] However, historians tend to view the Estado Novo as para-fascist in nature,[205] possessing minimal fascist tendencies.[206] Other historians, including Fernando Rosas and Manuel Villaverde Cabral, think that the Estado Novo should be considered fascist.[207]
Giorgio Almirante, leader of the Italian Social Movement from 1969 to 1987
The term neo-fascism refers to fascist movements after World War II. In Italy, the Italian Social Movement led by Giorgio Almirante was a major neo-fascist movement that transformed itself into a self-described "post-fascist" movement called the National Alliance (AN), which has been an ally of Silvio Berlusconi's Forza Italia for a decade. In 2008, AN joined Forza Italia in Berlusconi's new party The People of Freedom, but in 2012 a group of politicians split from The People of Freedom, refounding the party with the name Brothers of Italy. In Germany, various neo-Nazi movements have been formed and banned in accordance with Germany's constitutional law which forbids Nazism. The National Democratic Party of Germany (NPD) is widely considered a neo-Nazi party, although the party does not publicly identify itself as such.
In Argentina, Peronism, associated with the regime of Juan Perón from 1946 to 1955 and 1973 to 1974, was influenced by fascism.[208] Between 1939 and 1941, prior to his rise to power, Perón had developed a deep admiration of Italian Fascism and modelled his economic policies on Italian fascist policies.[208] However, not all historians agree with this identification,[209] which they consider debatable[210] or even false,[211] biased by a pejorative political position.[212] Other authors, such as the Israeli Raanan Rein, categorically maintain that Perón was not a fascist and that this characterization was imposed on him because of his defiant stance against US hegemony.[213]
Contemporary fascism (2008–present)
See also: Alt-right, Radical right (United States), and Fascism in the United States
Greece
Main article: Golden Dawn
Golden Dawn demonstration in Greece in 2012
After the onset of the Great Recession and economic crisis in Greece, a movement known as the Golden Dawn, widely considered a neo-Nazi party, soared in support out of obscurity and won seats in Greece's parliament, espousing a staunch hostility towards minorities, illegal immigrants and refugees. In 2013, after the murder of an anti-fascist musician by a person with links to Golden Dawn, the Greek government ordered the arrest of Golden Dawn's leader Nikolaos Michaloliakos and other members on charges related to being associated with a criminal organization.[214][215] On 7 October 2020, Athens Appeals Court announced verdicts for 68 defendants, including the party's political leadership. Nikolaos Michaloliakos and six other prominent members and former members of parliament (MPs) were found guilty of running a criminal organization.[216] Guilty verdicts were delivered on charges of murder, attempted murder, and violent attacks on immigrants and left-wing political opponents.[217]
Post-Soviet Russia
Main articles: Rashism and Putinism
Marlene Laruelle, a French political scientist, contends in Is Russia Fascist? that the accusation of "fascist" has evolved into a strategic narrative of the existing world order. Geopolitical rivals might construct their own view of the world and assert the moral high ground by branding ideological rivals as fascists, regardless of their real ideals or deeds. Laruelle discusses the basis, significance, and veracity of accusations of fascism in and around Russia through an analysis of the domestic situation in Russia and the Kremlin's foreign policy justifications; she concludes that Russian efforts to brand its opponents as fascist is ultimately an attempt to determine the future of Russia in Europe as an antifascist force, influenced by its role in fighting fascism in World War II.[218]
According to Alexander J. Motyl, an American historian and political scientist, Russian fascism has the following characteristics:[219][220]
An undemocratic political system, different from both traditional authoritarianism and totalitarianism;
Statism and hypernationalism;
A hypermasculine cult of the supreme leader (emphasis on his courage, militancy and physical prowess);
General popular support for the regime and its leader.[221]
Protester against the Russian government, holding an image portraying Dmitry Medvedev and Vladimir Putin as Nazis with a swastika made of colours of the Ribbon of Saint George and a Russian coat of arms in the centre (Odesa, 2014)
Yale historian Timothy Snyder has stated, "Putin's regime is ... the world center of fascism" and has written an article entitled "We Should Say It: Russia Is Fascist."[222] Oxford historian Roger Griffin compared Putin's Russia to the World War II-era Empire of Japan, saying that like Putin's Russia, it "emulated fascism in many ways, but was not fascist."[223] Historian Stanley G. Payne says Putin's Russia "is not equivalent to the fascist regimes of World War II, but it forms the nearest analogue to fascism found in a major country since that time" and argues that Putin's political system is "more a revival of the creed of Tsar Nicholas I in the 19th century that emphasized 'Orthodoxy, autocracy, and nationality' than one resembling the revolutionary, modernizing regimes of Hitler and Mussolini."[223] According to Griffin, fascism is "a revolutionary form of nationalism" seeking to destroy the old system and remake society, and that Putin is a reactionary politician who is not trying to create a new order "but to recreate a modified version of the Soviet Union". German political scientist Andreas Umland said genuine fascists in Russia, like deceased politician Vladimir Zhirinovsky and activist and self-styled philosopher Aleksandr Dugin, "describe in their writings a completely new Russia" controlling parts of the world that were never under tsarist or Soviet domination.[223] According to Marlene Laurelle writing in The Washington Quarterly, "applying the "fascism" label ... to the entirety of the Russian state or society short-circuits our ability to construct a more complex and differentiated picture."[222]
Radio Free Europe/Radio Liberty, collecting the opinions of experts on fascism, said that while Russia is repressive and authoritarian, it cannot be classified as a fascist state for various reasons, including Russia's government being more reactionary than revolutionary.[223] In 2023, Oleg Orlov, the chairman of the Board of Human Rights Center "Memorial", claimed that Russia under Vladimir Putin had descended into fascism and that the army is committing "mass murder".[224][225] On 7 March 2024, in his 2024 State of the Union Address, American President Joe Biden compared Russia under Vladimir Putin to Adolf Hitler's conquests of Europe.[226]
Tenets
Robert O. Paxton finds that even though fascism "maintained the existing regime of property and social hierarchy", it cannot be considered "simply a more muscular form of conservatism" because "fascism in power did carry out some changes profound enough to be called 'revolutionary.'"[227] These transformations "often set fascists into conflict with conservatives rooted in families, churches, social rank, and property." Paxton argues:
fascism redrew the frontiers between private and public, sharply diminishing what had once been untouchably private. It changed the practice of citizenship from the enjoyment of constitutional rights and duties to participation in mass ceremonies of affirmation and conformity. It reconfigured relations between the individual and the collectivity, so that an individual had no rights outside community interest. It expanded the powers of the executive—party and state—in a bid for total control. Finally, it unleashed aggressive emotions hitherto known in Europe only during war or social revolution.[227]
Ultranationalism
Part of a series on
Nationalism
Nation forming
Core values
Types
Organizations
Related concepts
icon Politics portal
vte
Ultranationalism, combined with the myth of national rebirth, is a key foundation of fascism.[228] Robert Paxton argues that "a passionate nationalism" is the basis of fascism, combined with "a conspiratorial and Manichean view of history" which holds, "the chosen people have been weakened by political parties, social classes, unassimilable minorities, spoiled rentiers, and rationalist thinkers."[229] Roger Griffin identifies the core of fascism as being palingenetic ultranationalism.[42]
The fascist view of a nation is of a single organic entity that binds people together by their ancestry and is a natural unifying force of people.[230] Fascism seeks to solve economic, political, and social problems by achieving a millenarian national rebirth, exalting the nation or race above all else and promoting cults of unity, strength, and purity.[231][page needed][232][9] European fascist movements typically espouse a racist conception of non-Europeans being inferior to Europeans.[233] Beyond this, European fascists have not held a unified set of racial views.[233] Historically, most fascists promoted imperialism, although there have been several fascist movements that were uninterested in the pursuit of new imperial ambitions.[233] For example, Nazism and Italian Fascism were expansionist and irredentist. Falangism in Spain envisioned the worldwide unification of Spanish-speaking peoples (Hispanidad). British Fascism was non-interventionist, though it did embrace the British Empire.
Totalitarianism
Fascism promotes the establishment of a totalitarian state.[13] It opposes liberal democracy, rejects multi-party systems, and may support a one-party state so that it may synthesize with the nation.[14] Mussolini's The Doctrine of Fascism (1932), partly ghostwritten by philosopher Giovanni Gentile,[234] who Mussolini described as "the philosopher of Fascism", states: "The Fascist conception of the State is all-embracing; outside of it no human or spiritual values can exist, much less have value. Thus understood, Fascism is totalitarian, and the Fascist State—a synthesis and a unit inclusive of all values—interprets, develops, and potentiates the whole life of a people."[235] In The Legal Basis of the Total State, Nazi political theorist Carl Schmitt described the Nazi intention to form a "strong state which guarantees a totality of political unity transcending all diversity" in order to avoid a "disastrous pluralism tearing the German people apart."[236]
Fascist states pursued policies of social indoctrination through propaganda in education and the media, and regulation of the production of educational and media materials.[237] Education was designed to glorify the fascist movement and inform students of its historical and political importance to the nation. It attempted to purge ideas that were not consistent with the beliefs of the fascist movement and to teach students to be obedient to the state.[238]
Economy
Main article: Economics of fascism
Historians and other scholars disagree on the question of whether a specifically fascist type of economic policy can be said to exist. David Baker argues that there is an identifiable economic system in fascism that is distinct from those advocated by other ideologies, comprising essential characteristics that fascist nations shared.[239] Payne, Paxton, Sternhell et al. argue that while fascist economies share some similarities, there is no distinctive form of fascist economic organization.[240][page needed][241][page needed][231][page needed] Gerald Feldman and Timothy Mason argue that fascism is distinguished by an absence of coherent economic ideology and a lack of serious economic thinking. They state that the decisions taken by fascist leaders cannot be explained within a logical economic framework.[242]
Fascists presented their views as an alternative to both international socialism and free-market economics.[243] While fascism opposed mainstream socialism, fascists sometimes regarded their movement as a type of nationalist "socialism" to highlight their commitment to nationalism, describing it as national solidarity and unity.[244][245] Fascism had a complex relationship with capitalism, both supporting and opposing different aspects of it at different times and in different countries. In general, fascists held an instrumental view of capitalism, regarding it as a tool that may be useful or not, depending on circumstances.[246][247] Fascist governments typically established close connections between big business and the state, and business was expected to serve the interests of the government.[246][247] Economic self-sufficiency, known as autarky, was a major goal of most fascist governments.[248]
Fascist governments advocated for the resolution of domestic class conflict within a nation in order to guarantee national unity.[249] This would be done through the state's mediating relations between the classes (contrary to the views of classical liberal-inspired capitalists).[250] While fascism was opposed to domestic class conflict, it held that bourgeois-proletarian conflict existed primarily in international conflict between proletarian nations and bourgeois nations.[251] Fascism condemned what it viewed as widespread character traits that it associated with the typical bourgeois mentality that it opposed, such as materialism, crassness, cowardice, and the inability to comprehend the heroic ideal of the fascist "warrior"; and associations with liberalism, individualism, and parliamentarianism.[252] In 1918, Mussolini defined what he viewed as the proletarian character, defining proletarian as being one and the same with producers, a productivist perspective that associated all people deemed productive, including entrepreneurs, technicians, workers and soldiers as being proletarian.[citation needed]
The need for a people's car (Volkswagen in German), its concept and its functional objectives were formulated by Adolf Hitler.
Because productivism was key to creating a strong nationalist state, it criticized internationalist and Marxist socialism, advocating instead to represent a type of nationalist productivist socialism. Nevertheless, while condemning parasitical capitalism, it was willing to accommodate productivist capitalism within it so long as it supported the nationalist objective.[253] The role of productivism was derived from Henri de Saint Simon, whose ideas inspired the creation of utopian socialism and influenced other ideologies that stressed solidarity rather than class war and whose conception of productive people in the economy included both productive workers and productive bosses to challenge the influence of the aristocracy and unproductive financial speculators.[254] Saint Simon's vision combined the traditionalist right-wing criticisms of the French Revolution with a left-wing belief in the need for association or collaboration of productive people in society.[254] Whereas Marxism condemned capitalism as a system of exploitative property relations, fascism saw the nature of the control of credit and money in the contemporary capitalist system as abusive.[253]
Unlike Marxism, fascism did not see class conflict between the Marxist-defined proletariat and the bourgeoisie as a given or as an engine of historical materialism.[253] Instead, it viewed workers and productive capitalists in common as productive people who were in conflict with parasitic elements in society, including corrupt political parties, corrupt financial capital, and feeble people.[253] Fascist leaders such as Mussolini and Hitler spoke of the need to create a new managerial elite led by engineers and captains of industry—but free from the parasitic leadership of industries.[253] Hitler stated that the Nazi Party supported bodenständigen Kapitalismus ("productive capitalism") that was based upon profit earned from one's own labour, but condemned unproductive capitalism or loan capitalism, which derived profit from speculation.[255]
Fascist economics supported a state-controlled economy that accepted a mix of private and public ownership over the means of production.[256] Economic planning was applied to both the public and private sectors, and the prosperity of private enterprise depended on its acceptance of synchronizing itself with the economic goals of the state.[186] Fascist economic ideology supported the profit motive but emphasized that industries must uphold the national interest as superior to private profit.[186]
While fascism accepted the importance of material wealth and power, it condemned materialism, which was identified as being present in both communism and capitalism, and criticized materialism for lacking acknowledgment of the role of the spirit.[257] In particular, fascists criticized capitalism, not because of its competitive nature nor support of private property, which fascists supported—but due to its materialism, individualism, alleged bourgeois decadence and alleged indifference to the nation.[258] Fascism denounced Marxism for its advocacy of materialist internationalist class identity, which fascists regarded as an attack upon the emotional and spiritual bonds of the nation and a threat to the achievement of genuine national solidarity.[259]
In discussing the spread of fascism beyond Italy, historian Philip Morgan states:
Since the Depression was a crisis of laissez-faire capitalism and its political counterpart, parliamentary democracy, fascism could pose as the 'third-way' alternative between capitalism and Bolshevism, the model of a new European 'civilization.' As Mussolini typically put it in early 1934, 'from 1929 ... fascism has become a universal phenomenon ... The dominant forces of the 19th century, democracy, socialism, [and] liberalism have been exhausted ... the new political and economic forms of the twentieth-century are fascist'.[260]
Fascists criticized egalitarianism as preserving the weak and instead promoted social Darwinist views and policies.[261][262] They were in principle opposed to the idea of social welfare, arguing that it "encouraged the preservation of the degenerate and the feeble."[263] The Nazi Party condemned the welfare system of the Weimar Republic, as well as private charity and philanthropy, for supporting people whom they regarded as racially inferior and weak and who should have been weeded out in the process of natural selection.[264] Nevertheless, faced with the mass unemployment and poverty of the Great Depression, the Nazis found it necessary to set up charitable institutions to help racially pure Germans in order to maintain popular support while arguing that this represented "racial self-help" and not indiscriminate charity or universal social welfare.[265] Thus, Nazi programs such as the Winter Relief of the German People and the broader National Socialist People's Welfare (NSV) were organized as quasi-private institutions, officially relying on private donations from Germans to help others of their race—although in practice those who refused to donate could face severe consequences.[266] Unlike the social welfare institutions of the Weimar Republic and the Christian charities, the NSV distributed assistance on explicitly racial grounds. It provided support only to those who were "racially sound, capable of and willing to work, politically reliable, and willing and able to reproduce." Non-Aryans were excluded, as well as the "work-shy", "asocials" and the "hereditarily ill".[267] Under these conditions, by 1939, over 17 million Germans had obtained assistance from the NSV, and the agency "projected a powerful image of caring and support" for "those who were judged to have got into difficulties through no fault of their own."[267] Yet the organization was "feared and disliked among society's poorest" because it resorted to intrusive questioning and monitoring to judge who was worthy of support.[268]
Direct action
Fascism emphasizes direct action, including supporting the legitimacy of political violence, as a core part of its politics.[269] Fascism views violent action as a necessity in politics that fascism identifies as being an "endless struggle";[270] this emphasis on the use of political violence means that most fascist parties have also created their own private militias (e.g. the Nazi Party's Brown shirts and Fascist Italy's Blackshirts). The basis of fascism's support of violent action in politics is connected to social Darwinism.[270] Fascist movements have commonly held social Darwinist views of nations, races, and societies.[271] They say that nations and races must purge themselves of socially and biologically weak or degenerate people while simultaneously promoting the creation of strong people in order to survive in a world defined by perpetual national and racial conflict.[272]
Age and gender roles
Members of the Piccole Italiane, an organization for girls within the National Fascist Party in Italy
Members of the League of German Girls, an organization for girls within the Nazi Party in Germany