1 Reprinted with Permission
2 -------------------------
3
4
5
6 The ARTFL Project Newsletter
7 Volume 8, Number 1 - Winter 1992-93
8
9 American and French Research on the
10 Treasury of the French Language
11
12 ARTFL is a cooperative project between:
13 Centre National de la Recherche Scientifique and
14 The University of Chicago
15
16
17
18 Particular Vices in Decent Expressions
19
20 Who uses this thing, anyway? And how?
21
22 Over the past years use of the ARTFL database has con-
23 tinued to increase steadily. From eight subscribing insti-
24 tutions in 1988, the ARTFL Consortium has grown to over
25 forty institutions in the United States and Canada. We are
26 beginning to see its impact on both teaching and scholar-
27 ship. Teachers and researchers turn to it to trace the
28 introduction and development of new concepts and the rework-
29 ing of old ones over time; to study the crystallization of
30 ideas in key terms; to examine patterns and shifts in
31 language use; to look at the dynamics at work between par-
32 ticular texts and more general usage; and, more simply, to
33 help students get a better sense of the language. The fol-
34 lowing articles give some examples of the ways scholars are
35 putting the ARTFL database to work in their research and in
36 the classroom. The methods and "ideologies" range widely;
37 from using philological techniques to studying intertextual-
38 ity, from demonstrating the originality of a writer to cele-
39 brating the death of the author.
40
41 I would like to thank the authors of these articles for
42 their contributions and invite other users to send to ARTFL
43 descriptions of how they have used the database in their
44 research and teaching. Understanding the ways in which peo-
45 ple use the database as well as the problems encountered in
46 working with it and ideas for different types of uses is
47 important for us in determining how the database should grow
48 and what new tools and access methods to develop.
49
50 Robert Morrissey, Director
51
52
53
54
55
56
57
58
59
60
61
62
63 December 2, 1992
64
65
66
67
68 - 2 -
69
70
71 Table of Contents
72
73 RESEARCH
74
75 Daniel Gordon, Governing Ideas: A Philological Approach
76 to the Age of Enlightenment
77 Keith Baker, Public Opinions and Revolutionary
78 Thoughts: Searching for Eighteenth-Century
79 Political Culture.
80 Geoffrey Wall, _Amboche, Masure, Nankins_: Coming to Terms
81 with _Madame Bovary_.
82
83 CLASSROOM
84
85 William Winder, _La puce a` l'oreille_: Enigmatexts at
86 the University of British Columbia
87 Michel Grimaud, Being Proper: ARTFL in Undergraduate
88 Teaching Wellesley College.
89 Jean-Claude Carron, The Nitty Gritty: ARTFL in the
90 Introduction to Literary Studies at UCLA
91 Announcements
92
93
94 RESEARCH
95
96 Governing Ideas: A Philological
97 Approach to the Age of Enlightenment
98
99 The ARTFL database has been a great resource for me in
100 my research on the moral and political vocabularies that
101 were invented in the Enlightenment. Works such as
102 Voltaire's _Dictionnaire philosophique_ and the
103 _Encyclope'die_ of Diderot and d'Alembert indicate that the
104 creation of a new moral and political lexicon was a central
105 concern of the philosophes. Rousseau wrote, "Every estate
106 and every profession has its own dictionary that defines its
107 particular vices in decent expressions." Rousseau and oth-
108 ers sought to replace inaccurate and deceitful language with
109 a new and enlightened terminology. Historical semantics is
110 thus a useful method for approaching the Enlightenment
111 because Enlightenment authors themselves viewed their role
112 in semantic terms. It is also a useful method for intellec-
113 tual history in general because it provides a concrete way
114 of referring both to intellectual innovation and to the com-
115 mon intellectual orientation of a large number of writers
116 over a long timespan.
117
118 The ARTFL database has been valuable to me as a tool
119 for tracing the usage of particular words that seem to have
120 been central in the preferred lexicon of the _philosophes_.
121 I have focused mainly on the language of politeness -- terms
122 such as _sociabilite'_, _civilite'_, and _politesse_ which
123 Rousseau used with great ambivalence but which most Enlight-
124 enment philosophers used with great conviction. The
125
126
127
128 December 2, 1992
129
130
131
132
133 - 3 -
134
135
136 database makes it possible to test hunches and hypotheses
137 efficiently. For example, the _Encyclope'die_ was the first
138 French dictionary to include the word _sociabilite'_. Does
139 this mean that the term had been recently coined and that
140 the editors of the _Encyclope'die_ were consciously trying
141 to confer legitimacy on a new word or concept? How does the
142 definition of _sociabilite'_ in the _Encyclope'die_ resemble
143 or differ from previous usages? Is it possible that the
144 term _socie'te'_ from which _sociabilite'_ derives, a term
145 that has become so ubiquitious that we can scarcely imagine
146 its non-existence -- is it possible that this term was also
147 put into circulation by Enlightenment authors? (This last
148 question, which was suggested to me by Keith Baker, has
149 become the basis of collaborative research.) By searching
150 the ARTFL database for occurences of _sociabilite'_ prior
151 to the publication of the _Encyclope'die_ and searching the
152 entire database for occurences of _socie'te'_, I have gained
153 information that allows these broad questions to be
154 approached precisely.
155
156 Recently, I came across an interesting remark of Lucien
157 Febvre, who was an advocate of the study of words. The
158 value of tracing the evolution of words, he wrote, is that
159 "they reach us pregnant, one might say, with all the history
160 through which they have passed. They alone can enable us to
161 follow and measure...the transformations which took place in
162 those governing ideas which man is pleased to think of as
163 being immobile because their immobility seems to be a
164 guarantee of his security." The ARTFL database has allowed
165 me to confirm that a number of terms which we utter uncons-
166 ciously were consciously invented and put into circulation
167 in the Enlightenment. (Daniel Gordon's _The Idea of Socia-
168 bility in Pre-Revolutionary France_ will be published by
169 Princeton University Press in 1993.)
170
171 Daniel Gordon
172 Harvard University
173
174
175 Public Opinions and Revolutionary Thoughts: Searching for
176 Eighteenth-Century Political Culture.
177
178 Over the years, I have used ARTFL in a number of
179 research projects on the history of French political cul-
180 ture. My use of the database has been relatively straight-
181 forward and unsophisticated, but I have found it extremely
182 helpful. Generally speaking, I have searched the database
183 for occurrences of terms relevant to particular political
184 concepts. The searches have helped me to identify works
185 relevant to my project that I would not have anticipated, as
186 well as making it easier to find key occurrences of terms in
187 works that were obviously relevant. They have demonstrated
188 shifts in the frequency of the uses of important terms in
189 the database over relatively long periods of time. They
190
191
192
193 December 2, 1992
194
195
196
197
198
199 - 4 -
200
201
202 have also suggested evidence for changing uses of these
203 terms within differing discursive configurations.
204
205 One of the earliest uses I made of the database was for
206 a study of the idea of "public opinion" in eighteenth-
207 century France, first published in 1987 and subsequently
208 reprinted in my work, _Inventing the French Revolution_
209 (Cambridge, 1990). At that time, it was not possible -- as
210 it now would be -- to search the database for co-occurrences
211 of the terms _opinion_ and _publique_. Instead, we had to
212 search for occurrences of _opinion_ more generally, and then
213 find out which of these occurrences actually combined _opin-
214 ion_ with _publique_. The procedure was somewhat cumber-
215 some, but it was enormously useful in identifying
216 occurrences of _opinion publique_ in the database for
217 further analysis, in suggesting a tentative chronology for
218 the usage of the term in eighteenth-century France, and in
219 illustrating the traditional associations of _opinion_ with
220 uncertainty, instability, and disorder -- associations that
221 were rapidly changed when mere _opinion_ was transformed (as
222 it was during the third quarter of the eighteenth-century)
223 into the rational authority of of _opinion publique_, the
224 new tribunal to which all political actors were compelled to
225 appeal.
226
227 Another project in which I had valuable recourse to
228 ARTFL was a study of the idea of "revolution" in prerevolu-
229 tionary France, first published in 1988 and also reprinted
230 in _Inventing the French Revolution_. Searching the database
231 for _re'volution_ produced an enormous amount of informa-
232 tion. It revealed important occurrences in works I would
233 not otherwise have investigated, as well as ensuring that I
234 did not miss occurrences in works I already knew to be cru-
235 cial (Mably's _Observations sur l'histoire de France_, for
236 example). It also provided the basis for the following
237 table, adapted from _Inventing the Revolution_, p. 346:
238
239 Frequency of occurrences of _re'volution(s)_
240 in the ARTFL database (1986)
241
242
243 Date Number of Number of words Frequency
244 per
245 occurrences in corpus 1,000 words
246
247 1600-99 152 18,269,513 .0083
248 1700-99 2,526 37,499,880 .0673
249
250 1700-50 392 12,805,037 .0306
251 1751-70 782 10,879,911 .0718
252 1771-89 504 10,651,996 .0473
253 1789-99 848 3,162,936 .2681
254
255 Of course, as I pointed out in presenting the table, the
256 ARTFL database is not, in any strict statistical sense, a
257
258
259
260 December 2, 1992
261
262
263
264
265 - 5 -
266
267
268 representative sample of French works published during the
269 seventeenth- and eighteenth-centuries. One cannot therefore
270 extrapolate directly from the frequency of the term
271 _re'volution_ in the database to its popularity in French
272 discourse as a whole. Nonetheless, the increase of
273 occurrences of the term within the database, between the
274 seventeenth- and eighteenth-centuries, is really quite
275 striking. It does suggest, along with other evidence, that
276 "revolution" was becoming an increasingly important category
277 of historical understanding in France well before the events
278 that were so quickly apprehended in 1789 as "The French
279 Revolution." I also tried in my study to identify the com-
280 peting meanings of _re'volution_ in prerevolutionary politi-
281 cal discourse and to show how the French Revolution was
282 invented as a series of improvisations upon them. More
283 recently, Daniel Gordon and I have been collaborating on
284 research in the idea of society in the French Enlightenment.
285
286 I have learned much from ARTFL and expect to be able to
287 learn much more in the future. Allow me to conclude by
288 thanking the ARTFL staff for their generous help over the
289 years, and to acknowledge the particular assistance of Kent
290 Wright and Matthew Levinger in the searches they have so
291 readily carried out for me.
292
293 Keith Baker
294 Stanford University
295
296
297 _Amboche, Masure, Nankins_: Coming to
298 Terms with _Madame Bovary_.
299
300 After four years' work, I have just finished a new
301 translation of Flaubert's _Madame Bovary_. As well as the
302 usual array of translator's tools--the various dictionaries,
303 the thesaurus and the encyclopedia, I have been using ARTFL.
304
305 It has helped me in three ways.
306
307 First of all, it meant that I could easily generate a
308 voluminous file of quotations systematically, illustrating
309 the semantic field of key words in Flaubert's text. For
310 example, I searched the fiction and the correspondance for
311 instances of "nerfs" and "nerveux": an obvious case of a
312 word which carries a special thematic charge for Flaubert.
313 There were less obvious instances too, such as "abandonner,"
314 "langueur," and "songer." All of these were used quite dis-
315 tinctively. ARTFL supplemented and refined my intuitive
316 sense of their meaning, helped me to make confident and con-
317 sistent decisions in my translation of recurrent words.
318
319 Second, there were other kinds of words, such as
320 "amboche," "masure" and "nankins." They raised different
321 problems: of connotation, of cultural history. The
322
323
324
325 December 2, 1992
326
327
328
329
330
331 - 6 -
332
333
334 dictionaries and the encyclopedias didn't give enough exam-
335 ples of their use to answer the questions I was asking.
336 ARTFL could do it, rapidly and exactly.
337
338 Third, when writing my introduction I wanted to say
339 something about (for example) Flaubert's opinions on hys-
340 teria. ARTFL confirmed my intuition that the word "hys-
341 teria" did not occur in the text of _Madame Bovary_. And it
342 revealed, simultaneously, how much thought Flaubert had
343 given to the question of hysteria, how informed he was con-
344 cerning the emergent domain of psychopathology.
345
346 I think we are just learning how to use ARTFL as an
347 instrument of research. Our intellectual formation in the
348 pre-electronic era means that we still do not, in general,
349 try to follow up any of those criticial intuitions that
350 would once have required weeks of drudgery to substantiate.
351 We are learning, belatedly, to ask the right questions.
352 (Geoffrey Wall's translation of _Madame Bovary_ was pub-
353 lished by Penguin Books in September, 1992.)
354
355 Geoffrey Wall
356 University of York
357
358
359
360 CLASSROOM
361
362
363 _La puce a` l'oreille_: Enigmatexts at
364 the University of British Columbia
365
366 French 500 is a methodology and bibliography course
367 offered to beginning masters and doctoral candidates at
368 UBC's French Department. Much of the course is devoted to
369 the traditional questions, procedures, and tools of humani-
370 ties graduate work, i.e. thesis format and goals, the
371 library and its printed reference materials, and terminolog-
372 ical and methodological conventions for linguistic and
373 literary studies. Students are also presented a variety of
374 electronic research tools: electronic mail, electronic dis-
375 cussion lists, optical scanners, word processors, biblio-
376 graphical databases, terminological databases, text analysis
377 software, grammar correction software, etc.
378
379 Among these electronic tools, the ARTFL database is
380 perhaps the most important for French studies. Accordingly,
381 students are asked to complete a number of practical exer-
382 cises in order to gain a more thorough understanding of its
383 uses and limits. Some examples of exercises are:
384
385 Use ARTFL to compare the value of "e^tre paru" and "avoir
386 paru" in the sense of "published": is there any difference
387 in meaning?
388
389
390
391 December 2, 1992
392
393
394
395
396
397 - 7 -
398
399
400 Study the evolution of the fixed expression "puce a`
401 l'oreille": how is it used over the centuries and what is
402 its syntactic distribution?
403
404 Compare the role of "vin" and "champagne" in Maupassant and
405 Zola. What are the connotations of these words, and what
406 role do they have in the text?
407
408 Construct a thematic field for Flaubert's "Un coeur simple";
409 does this field appear in other works?
410
411 Finally, perhaps the most ambitious exercise of the
412 course consists in finding the source of unidentified quota-
413 tions of varying length, called "enigmatexts". Students are
414 asked to find the exact edition that the quotation was taken
415 from, using whatever means possible. ARTFL is one particu-
416 larly effective tool for this exercise, but not the only
417 one, and it is often instructive to compare the different
418 approaches to the problem. For example, faced with the unti-
419 tled passage "Par la fene^tre losangique d'une gui^rite de
420 gardien, je suivais le chapeau panama dans les alle'es du
421 Luxembourg", one student judiciously looked in the _TLF_
422 under "losangique" and found that passage as an attestation.
423 Finding this double path to the text served as an excellent
424 illustration of the circulation of information between
425 reference texts and primary sources: such explorations
426 reveal the dictionary-like dimension of textual databases,
427 and the textual dimension of the dictionary.
428
429 Some enigmatexts cannot be found with ARTFL, either
430 because the texts are not part of the database or because
431 the relevant search is not possible. For instance, any pas-
432 sage that is composed exclusively of high frequency words,
433 such as grammatical words ("de", "par", "le", etc.), is dif-
434 ficult if not impossible to find in ARTFL. Indeed ARTFL gen-
435 erally suffers from a lack of statistical and grammatical
436 treatment of its texts -- for example, one cannot use it to
437 study in any direct way the syntax of determiners or the
438 statistical distribution of a theme in a text. However,
439 these limitations are themselves pedagogically valuable in
440 that they serve as a natural transition to the more complex
441 questions of automatic treatment of text, and to the use of
442 more powerful software for linguistic and literary study:
443 automatic lemmatizers, concordances, linguistic databases,
444 etc.
445
446 In the final stages of the course, students create a
447 miniature database from Flaubert`s _Un Coeur Simple_. They
448 draw up extensive lists of expressions and phrases from the
449 novella and then search the whole database for significant
450 parallels. This very practical study of intertextuality
451 helps students to understand that beyond the simple biblio-
452 graphical attribution, the true source of the language,
453 style, and meaning of texts is found dispersed in the whole
454
455
456
457 December 2, 1992
458
459
460
461
462
463 - 8 -
464
465
466 of literature itself -- in a very real sense Flaubert did
467 not write "Un coeur simple", rather French literature did.
468 ARTFL is perhaps the only reference tool that gives direct
469 access to that elusive, polycephalic author, French literary
470 history.
471
472 Bill Winder
473 University of British Columbia
474
475
476 Being Proper: ARTFL in Undergraduate
477 Teaching Wellesley College.
478
479 Students in an undergraduate seminar on "Politeness and
480 Proper Names" have used the ARTFL database for their papers.
481 One student studied the use of "tu" and "vous" in Balzac's
482 _Euge'nie Grandet_, noting for example Grandet's switch from
483 "tu" to "vous" when he felt emotionally distanced from his
484 wife or from Euge'nie. Another student looked at the use of
485 titles of address in Balzac's _Pe`re Goriot_, examining in
486 particular the use of "Mademoiselle" and "Madame" and, for
487 instance, the logic of variations between "Victorine,"
488 "Mademoiselle Taillefer," and "Mademoiselle Victorine."
489
490 One student looked at Simenon's _Les vacances de
491 Maigret_, also studying the rare cases of "tutoiement" and
492 also the use of the article with a given name as in "la
493 Popine" vs. "Popine."
494
495 Points of convergence between Simenon and Balzac were
496 easy to investigate thanks to the ease of ARTFL searches: we
497 checked the now odd use of "Mme + Patronym" as in "Mme
498 Maigret" said by husband to wife and found it, for instance,
499 in Balzac's _Euge'nie Grandet_.
500
501 A study of naming in Jules Verne's _Ile myste'rieuse_
502 was particularly interesting in several ways. We looked at
503 the use of "Monsieur" with a given name for men [cf. Balzac
504 above for women] as in "Monsieur Cyrus" vs. "Monsieur
505 Smith." We also discussed the usual use of the full name by
506 the narrator -- "Cyrus Smith," a way of naming also used by
507 Victor Hugo in _Les Miserables_ where Hugo always says "Jean
508 Valjean" (never "Jean" or "Valjean") In addition, KWIC lists
509 of names enabled us to study the order of mention of names
510 of the members of that small, strongly hierarchical social
511 group.
512
513 Many of the points discussed by the students were first
514 discussed in my 1989 article in _Le francais moderne_ ("Les
515 appellatifs dans le discours"). But, at the time, I did not
516 have access to the ARTFL database. Particularly in the last
517 case mentioned -- that of the order of mention of characters
518 -- having access to complete lists enabled the student who
519 studied the issue to significantly refine my initial
520
521
522
523 December 2, 1992
524
525
526
527
528
529 - 9 -
530
531
532 hypotheses.
533
534 Michel Grimaud
535 Wellesley College
536
537
538 The Nitty Gritty: ARTFL in the Introduction
539 to Literary Studies at UCLA.
540
541 The introductory course in literary studies, French
542 201: Literary Research and Composition, which is required
543 for all graduate students in the French Department at UCLA,
544 includes some information concerning the role and use of
545 computers in their graduate work. A certain number of these
546 uses begin as soon as the students realize that the catalo-
547 gue of the University's principal library is computerized,
548 as well as some ten years of the MLA Bibliography.
549
550 As far as ARTFL is concerned, the first "theoretical"
551 introduction is done in class: a description of the holdings
552 on the database, ways of accessing it (by modem, MOPS, or
553 "direct" connection), available resources (limiting the
554 corpus to areas of interest, breadths and limits of
555 searches, combinations of successive searches, etc.), possi-
556 ble uses (statistical, thematic, historical (first use of a
557 word), linguistic, phonetics and rhyme schemes, etc).
558
559 For those who have never used a computer before, there
560 is also the need to overcome intimidation and the numerous
561 unexpected difficulties raised by the initial encounter with
562 computer technology. After the first overview, the class
563 meets in a room equipped with a terminal that provides
564 access to UCLA's electronic mail and Bitnet via modem and
565 cable. (The Department does not yet have direct access to
566 Chicago.) The demonstration consists in part of an intro-
567 duction to communication programs compatible with our elec-
568 tronic service with modem access to Chicago. The use of the
569 database then happens "live", via modem, as does the stu-
570 dents' training in the use of Philologic. This is in fact
571 the only way for the class to see the program work and to
572 follow the advances, the experiments or the errors in the
573 research itself.
574
575 For their private research, the students are encouraged
576 to use MOPS which makes use of electronic mail.
577
578 Once these steps have been accomplished, access to
579 ARTFL is left relatively open: only the students who want to
580 use it or whose written course work might require it are
581 truly encouraged to use ARTFL for this introductory class.
582 French 201 offers essentially a series of introductions to
583 the different resources available to students for their
584 short term and long term research. The goal is mostly to
585 make them conscious that these resources exist, with the
586
587
588
589 December 2, 1992
590
591
592
593
594
595 - 10 -
596
597
598 hope that they will know how to use them when the time
599 comes.
600
601 For this first course, the numerous practical, adminis-
602 trative and theoretical steps seem to be enough to wear out
603 the curiosity and the patience of most of the students who
604 have never before used a computer. It is important to bear
605 in mind that this first step, though still not very
606 advanced, is crucial, and that we must not hurry things too
607 much.
608
609 Jean-Claude Carron
610 UCLA
611
612
613 ARTFL at the MLA
614
615 Mark Olsen, Assistant Director of ARTFL, will be at the
616 Modern Language Association. He will be acting as a respon-
617 dant for two sessions entitled "Signs, Symbols, Discourses:
618 A New Direction for Computer-Aided Literature Studies". The
619 sessions will be offered on Tuesday, December 29, (1:45-
620 3:00, Morgan Suite A&B, New York Hilton) and Wednesday,
621 December 30, (3:30-4:45, Riverside Ballroom, Sheraton New
622 York).
623
624 Mark will be happy to give ARTFL demonstrations by
625 appointment. Anyone interested in an ARTFL demonstration
626 should contact Mark to make an appointment either at the
627 conference or by e-mail to:
628 mark@gide.uchicago.edu,
629 or by phone at: 312-702-8687.
630
631
632 Morphological Analysis of the ARTFL database
633
634 We are pleased to announce that ARTFL now supports lim-
635 ited use of the INFL Morphological Analyzer under an agree-
636 ment with its developers at the Xerox Palo Alto Research
637 Center (PARC). The INFL Analyzer is a context-free system
638 which identifies many aspects of every word in a sentence,
639 including the tense, gender, part of speech, and other data.
640 The current implementation, which is not part of the Philo-
641 Logic access program, permits the user to search for a word
642 or pattern in a single text, generating a full morphological
643 analysis for every sentence in which the target word is
644 found. The INFL Analyzer will be discussed in further
645 detail in the Spring 1993 _Newsletter_. Please contact
646 ARTFL for further information concerning the use of INFL.
647
648
649 Current Subscriber List for ARTFL
650
651 We are pleased to announce that as of November, 1992,
652
653
654
655 December 2, 1992
656
657
658
659
660
661 - 11 -
662
663
664 thirty-five institutions are currently active participants
665 in ARTFL. These institutions are:
666
667 Bryn Mawr College
668 Carleton University
669 City University of New York
670 Columbia University
671 Cornell University
672 Dartmouth College
673 Duke University
674 Emory University
675 Gettysburg College
676 Harvard University
677 Indiana University
678 Johns Hopkins University
679 Kent State University
680 Louisiana State University
681 Memorial University
682 New York University
683 Princeton University
684 Rutgers University
685 Stanford University
686 State University of New York - Albany
687 State University of New York - Binghampton
688 State University of New York - Buffalo
689 Swarthmore College
690 Universite' de Montre'al
691 Universite' du Que'bec a` Montre'al
692 University of Alberta
693 University of British Columbia
694 University of California - Los Angeles
695 University of Illinois - Urbana-Champaign
696 University of Iowa
697 University of Manitoba
698 University of Michigan
699 University of Ottawa
700 University of Rochester
701 University of South Carolina
702 University of Southern California
703 University of Toronto
704 University of Virginia
705 University of Waterloo
706 Vassar College
707 Wellesley College
708 Yale University
709 Yeshiva University
710
711
712
713 ARTFL Project
714 Department of Romance Languages and Literatures
715 1050 East 59th Street
716 University of Chicago
717 Chicago, IL 60637
718
719
720
721 December 2, 1992
722
723
724
725
726
727 - 12 -
728
729
730 (312) 702-8488
731 artfl@artfl.uchicago.edu
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 December 2, 1992
788
789
790
791 <ONLINE MODERN HISTORY REVIEW>
This document was reprinted without modification.