Computers and the Humanities, Vol. 11, pp. 369-380. PERGAMONPRESS, 1977. Printed in the U.S.A.

Reviews Index des Regulae ad directionem ingenii de RenO Descartes, by J.-R. Armogathe and J.-L. Marion. Rome: Edizioni dell' Ateneo, 1976. Pp. xxii-163. By John E. Clark The short introduction to this important Descartes index rather uncharacteristically shows a greater sense of wonder at the international human effort that has produced it than at the use of a machine and a program. Emanating from the Equipe Descartes of the Centre National de la Recherche Scientifique in Paris, this first of a series of indexes of Descartes' works has used the computer and programs of the Laboratoire d'Analyse Statistique des Langues Anciennes (L.A.S.L.A.) in Liege, Belgium, and the printing has been assured by the Centro per un Lessico Intellettuale Europeo in Rome and the Consiglio Nazionale delle Ricerche. The format of this volume is simple. The AvantPropos (xi-xvi) has a first section signed by J.-R. Armogathe which states the problems encountered in setting up the Index and an explanation of the principles which allowed for the solutions. The lemmas of the Index are based on the Egidio Forcellini Lexicon totius latinitatis in the Padua (Corradini) edition of 1871, but many points of the classification of homographs apparently required "une r~gle plus rigoureuse" than Forcellini provided. Homographs gave the researchers, and will give the user of the Index, the biggest problem in what is otherwise a very clear presentation. When looking up, for example, a word like ABSOLVTVS, we immediately have the information that it is used eighteen times in Descartes' Regulae, and these 18 uses are neatly set out for us according to the various forms of gender, case and number with references to the page, line and word number in the line of the particular reference edition. Only a context is lacking; the text must be referred to. In the case of homographs, however, distinctions of meaning and grammatical function are made by

the use of a number following the lemma. While one can be only grateful for this sorting-out, the numbers used are anything but clear since they have not been assigned invariable meanings but have a particular meaning in each case. Generally, the numbers assigned refer to the order of grammatical functions (noun, adjective, numeral, adjective-pronoun etc.), and for each word the number I will refer to the first function in this list which applies. Thus TEMPVS I does not have the same meaning as VEL VT I. In addition, the change of number does not always indicate an absolute change in grammatical function. Other distinctions are coded by these same few numbers; some (for example VT) are explained in the Avant-Propos, which does not appear to have been written specifically for the Regulae (a word like MOROR is used as an example and it does not appear in this particular text). We must unfortunately conclude that the problem of homographs has not been solved in an ideal way. The second part of the Avant-Propos is signed by G. Crapulli and deals with the choice of the Adam and Tannery edition of the Descartes text (published in 1908 and reprinted in 1966) as the reference used in this Index. This is still the most important edition of Descartes' works although in individual cases it has been superseded. Mr. Crapulli himself published a better edition of the Regulae in 1966, from which he has contributed to the Index the fruit of his researches: the failings of the Adam-Tannery edition are clearly rectified by the pages immediately following the Avant-Propos ("Conjectures de Leibnitz," p. xviii, "Leqons omises dans le texte d'Adam-Tannery," p. xxii) and the corrections to the text by AdamTannery are also listed. The Index occupies almost the totality of the volume (pp. 1-149). The lemmas and their ordering are based, it will be recalled, on Forcellini; the precise form of the word in Descartes' text is indicated beneath the lemma. The presentation and information is thus, apart from the homograph distinctions, complete and clear. The volume concludes (pp.

0010-4817/77/1101-0369502.00/0 Copyright 9 1978 Pergamon Press

370

REVIEWS

153-163) with a "Liste des fr6quences," the words given in descending order of frequency. It is amusing to note that the lemma SVM has the greatest number of listings (724), although the form sum occurs only once. Very few words have not been given their complete list of occurrences (AB, AD, DE, ET 2, ETIA'II, EX, Lu NON, SED, I/EL 2, VT 4). This Index to Descartes' Regutae must be received with gratitude: a most useful tool for the study of an essential text, it appears to have been prepared with great care. Only one error has turned up after considerable spot-checking (VBI 5 should presumably be VBI 1, p. 142).

John E. Clark is an associate professor in the Department of French and Spanish at the University of Manitoba. He has recently published E16gie: The Fortunes of a Classical Genre in Sixteenth-Century France and completed research on Maro t's @ftres.

International Computing Symposium 1975: Proceedings of the International Computing Symposium 1975, Antibes, France, 2-4 June 1975. Ed. E. Gelenbe and D. Potier. New York: American Elsevier, and Amsterdam & Oxford: North Holland Publishing Company, 1975. Pp. viii + 266. $27.50, Dfl 75.00. ISBN 0-444-10901-3 (American Elsevier) or ISBN 0-7204-2839-4 (North-Holland). By R L Widmann This volume of thirty-six papers, selected from the more than one hundred and forty given at the ACM meeting in Antibes in 1975, offers a great variety of information and reflects a relatively high level of competence on the part of the various authors. The editors are to be congratulated for producing a book with good, clear, intelligible abstracts and for getting their authors' contributions to the readers in a format possible to follow. The double-columned pages have an open format with ample white space; illustrations and figures are easily read. The range of papers is enormous. Some concentrate on highly technical or theoretical examinations of problems, such as J. Kittler's "A Comparative Study of Five Locally Sensitive Clustering Techniques"; David L. Parnas, John E. Shore, and W. Dawd Elliott's "On the Need for Fewer Restrictions in Changing Compile-Time Environments"; and papers on aspects of numerical analyses by Bertrand Mercier, Bo Einarsson, and Daniel Gabay. Of more interest to the readers of CHum, most

probably, are the contributions describing the progress of several projects. P. A. Fortier discusses work at the University of Manitoba which produces Braille texts for students. The two programs devised by Fortier and tus colleagues, Don Keeping and J. Colin McConnell, have immediate results for blind students and also carry a wider social application, for the resulting texts are deposited m a central Braille library, available to all Canadian blind persons. This project, more than most others in the volume, illustrates and explains the connection between work in the academy and its social consequences. Fortier also makes clear the importance of a body like the Canada Council, which, in funding earlier projects of his, provided very important "seed money" for his literary research and thus allowed him to be instrumental in this project. Tire contribution of H. Zimmerman, "Terminal Access in the Cyclades Computer Network," deals with the problem of standards in the way terminals will work in computer networks. Though his outlook is not entirely sanguine, Zimmerman gives good evidence of a thoughtful, intelligent appraisal of some fundamental issues. Equally compelling, if not as well written, is Harald Geist's "Communication by Exception and a Tool to Achieve It in Man-Computer Interaction." Geist proposes to help fill the gap between batch and interactive computing for the terminal user by changing "the behaviour of the system." A brief and intriguing description of "A Structure-Oriented Program Editor: A First Step Towards Computer Assisted Programming" by V. Donzeau-Gouge, et al. at the Iria-Laboria makes some cogent and eminently sensible remarks on their project, which is programmed in PASCAL. Two interesting papers on developments in ALGOL 68 are "On Implementation of an ALGOL 68 Sublanguage" by L. Ammeraal and "Applications Support by Direct Language Extension-An Arbitrary Precision Arithmetic Facility in ALGOL 68" by J. L. Schonfelder and J.T. Thomason. A thoughtprovoking paper, "'Storage Design for Information Retrieval: Scarrott's Conjecture and Zipf's Law" by John M. Bennett, uses the analysis of 935 requests in the MEDLARS system at the University of Sydney (196%1971) to pose some general problems on storage technology. The volume has few typographical errors, but it lacks an index. It does, however, provide a useful list of the referees for ICS 1975 with their addresses.

R L Widmann, an associate professor of English at the University of Colorado at BouMer, is book review editor for CHum.

REVIEWS

Dynamic Information and Library ProcesshTg, Gerard Salton. Englewood Cliffs, N.J.: Prentice-Hall, 1975. Pp. xiv + 523. ISBN 0-13-221325-7. By Jean Tague Librarians have been confronted by computers, programmers by libraries. In the quest of one group for new tools and the other for new applications, a need has emerged for a comprehensive text about automated document processing. In three sections, Professor Salton's book fills this gap. Section I, "'State of the Art," is introductory, and places the reader in a library environment of automated acquisition, catalog support, circulation, indexing, abstracting, and retrieval systems. Section II, "Systems Analysis and Evaluation," requires greater mathematical diligence from the reader as it presents the models and evaluation measures currently of interest to information scientists. Section III, "Dynamic Information Processing," considers in greater depth the areas of file organization, automatic classification (clustering), language analysis, and interactive retrieval systems. Dynamic Information and Library Processing will be of more general interest than some of the author's earlier books. Although it describes his own work in computer-based document processing (the SMART and derivanve systems), it surveys, in addition, much of the other work in this field and includes extensive bibliographies. The emphasis is on dynamic systems, that is, systems which change in response to new input, both in the form of new documents and of user queries and relevance judgments. Much of the processing described in this book is experimental in the sense that it has been carried out with small files in a non-library setting. As the author himself points out, even the most effective automated processes will have little practical use until the machine-input problem is solved. Despite this limitation, the results are encouraging for those hopeful that automated processing can solve at least some of the problems of bibliographic control. Basing his comparisons primarily on the recall-precision curve,* the author has found that in many cases, retrieval systems based on machine indexing and classifying can outperform manual ones. The book has some omissions. More space might have been given to large-scale operational reference retrieval systems such as ORBIT, DIALOG, OCLC, and the New York Times Data Bank. Hardware *Recall is the proportion of the relevant documents which are retrieved; precision is the proportion of the retrieved documents which axe relevant.

371

requirements receive limited attention. The concept of computer-aided, as opposed to fully automated processing is becoming popular, as is exemplified by the widespread mterest in PRECIS. Yet it is barely mentioned. In general, though, these deficiencies appear to derive from the author's particular range of interests rather than any deliberate desire to ignore important aspects of the field. There is a useful distinction between prooforiented and insight-oriented procedures in the chapter on system testing. The succeeding review of retrieval system evaluations underlines, by their absence, the need for rigorous selection among the many insights of retrieval system designers and theoreticians. The author's espousal of quantification and exactness in the retrieval field has been a strong impetus in this needed direction. The material on file organization can be found in other computer science texts, but without the orientation to library and information center systems, as opposed to those for business. Data structures such as trees are mentioned, but are not formally defined or studied. Of particular interest to information system designers is the "clustered" file organization developed by the author. Detailed algorithms for clique and singleqink clustering are given, together with some advice on the problems and pitfalls of these and other clustering methods. The major obstacles to computer understanding of natural language are well-delineated. The novice reader might have found useful a glossary of such terms as syntactics, semantics, pragmatics, parsing, and synthetic relattons. At the moment, there is still some doubt concerning the value, in referenceretrieval as opposed to question-answering systems, of specifying syntactic relations. Salton presents some interesting empirical effects of varying the size of the indexing vocabulary on retrieval system performance. For terms within a document, discrimination value is defined as the difference between the compactness of the document space with and without the given term. Compactness, which is a function of inter-document similarity values, is thus reduced by selecting terms with a high discrimination value. From the results given, it appears that high discrimination values are associated with a high total term token to total document frequency. One hopes to see these investigations continued with larger collections. Query-based methods of classifying and indexterm weighting were found to rate highly on both effectiveness and efficiency measures. These results emphasize the importance of accumulating user queries and relevance judgments in order to optimize

372

REVIEWS

retrieval system performance. Unfortunately, the latter are almost impossible to collect on a regular and reliable basis m an operational setting. In summary, this book provides an excellent introduction to the experimental frontier in computer-based document processing. The emphasis is on the models and the algorithms themselves, rather than the computer/user interface problems or the wider question of the social implications of introducing into libraries energy-intensive processes to replace labor-intensive ones. To this particular focus, however, it brings a much-needed clarity and order.

Jean Tague, an associate professor in the School of Library and Information Science at the University of Western Ontario, has research interests in userresponsive bibliographic control and automation of community information centers.

R edeemers, Bourbons & Populists: Tennessee, 1870-1896. Roger L. Hart. Baton Rouge: Louisiana State University Press, 1975. By Philip R. VanderMeer Roger Hart's Redeemers, Bourbons & Populists is part of a growing re-examination of late nineteenthcentury politics. Like much of this literature it is quantitative, analyzing popular voting behavior as well as roll-call voting in state legislatures. Hart's major hypothesis is that Tennessee politics were shaped not by economic class conflict but by status anxiety. He thus seeks to replace the interpretations of C. Vann Woodward and Dan Robison with an approach which resembles that of his mentor, Sheldon Hackney. 1 Hart begins (somewhat abruptly) with the state constitutional convention of 1870, when Redeemers gained control of the state. Modifying Woodward's view, he shows that not all Redeemers were Whigindustrialists, nor did all Whig-industrialists become Redeemers. Controversy over the state debt ultimately destroyed the Redeemer coalition. Its firm support for full funding led in 1880 to a separate low-tax gubernatorial candidate, resulting in Republican victory. By 1882, aided by the defection of "Sky-Blues," the elite who demanded full funding, "Bourbons" controlled both party and state. Favoring compromise on the debt, limited government, a low tariff, and railroad regulation, Bourbons "combined postwar sectional resentment with the ancient Democratic

suspicion of industry and eastern cities" (p. 57). Seeming secure, Bourbon hegemony was soon disturbed by two forces. One was Governor Bob Taylor, first elected in 1886. Hart, stressing that Taylor was not a precursor of Populism, shows his support of New South men and measures, and describes his efforts as "the politics of style." A second challenge to the Bourbons was agrarian, beginning as the Farmer's Alliance and ultimately developing into Populism. In excluding merchants, lawyers, and bankers, "the Alliance defined its membership not by economic class, but by social status" (p. i 15). It was in large part the result of alienation. Opposed to the town-centered elite, it drew strength from the feeling that farmers "were not given the respect they deserved" (p. 115). Although initially nonpolitical, the movement soon abandoned this stance. In 1888 almost a third of the state legislators were Alliancemen, and in 1890 the Alliance President was elected governor. Insurgency was thus fairly successful, but as Hart demonstrates, Alliance legislators failed to vote as an anti-Bourbon bloc. After losing control of the Democratic party in 1892, some Alliancemen supported the Populist party, but only a few Populists were elected to the legislature. Hart interprets their occasional cooperation with Republicans as simply negative: "Populists were less interested in programs than in revenge" (p. 196). While Populism had somewhat greater appeal to poorer farmers, there were stronger connections "to frustrated expectations of upward social mobility or to the anomie of unstable growing communities" (p. 226). Redeemers, Bourbons & Populists is an interesting combination of several styles. As a political narrative it relies on an extensive reading of major newspapers, public documents, and private papers, and it covers in great (occasionally excessive) detail the state conventions, dealings of major politicians, and editorial opinion. As a quantitative study, it considers popular voting behavior as well as roll-call voting in the legislature and Democratic state conventions. Although there are maps, charts, and over 60 tables, the data analysis is not complex or difficult. Many of the tables are simple cross-tabulations, and the statistics used are multiple regression and phi coefficients. Such an approach is easy for laymen, but it also means that analysis is sometimes not pushed far enough. There are several technical problems, such as the use of a relatively low confidence level (.90), the almost-exclusive reliance on country level data, and the unclear handling of changing county boundaries. These flaws are not crucial, however, and the statistical analysis is generally competent.

REVIEWS The major problems with this book are not methodological but conceptual. Several are related to a general question of who the book is about. According to the subtitle and by constant implication, the subject is Tennessee politics. However, Republicans are seldom mentioned and never analyzed. This is a critical failure, for according to a variety of measures, about 45% of the electorate voted throughout this period for the GOP. Moreover, the omission leads Hart to misinterpret suffrage restriction, a discussion of which is unwisely relegated to an appendix. The work of J. Morgan Kousser offers an essential corrective on these points. 2 Even granting that the book deals only with Democrats, the subject is still unclear. Hart often discusses voters, party notables, or public officials without sufficient regard for their different political roles and characteristics or the varying quality of information about them. The similarities or differences among leaders of the Alliance, Bourbons, "Sky-Blues," or Redeemers are unclear, since the term "leaders" is neither defined nor analyzed systematically. Consequently, the discussion of factions, while detailed and complex, is inconclusive and somewhat confusing. Essentially the same problems plague the treatment of social status and the Farmer's Alliance. To show the political domination by a socio-economic elite, Hart mentions the offices and wealth or family of several individuals in three counties; to gauge the wealth of Alliancemen he calculates the landholding of twenty-five members located in three counties. The interpretation is certainly plausible, being congruent with recent works on Populism, but the evidence amounts to illustration rather than proof. 3 Some overlap of elites is to be expected, but how much overlap was there? How tightly-knit, or how open, was the elite? Hart does not answer these questions, nor does he test his thesis about the Alliance in politics. Were such men different than regular Democrats? Equally important, did the composition of membership or leadership change over time? Hart's emphasis on the power orientation of politicians often leads him to perceptive conclusions, but it also causes him to oversimplify. The state debt was more than symbolic. Moreover, as the voting data suggests, it was probably related not simply to farmer discontent but also to resentment of Republican railroad policies during Reconstruction. To praise as astute political judgment the Bourbon support for a low debt settlement ignores the fact that Bourbon leaders and followers were low-taxers. The relatively

373

easy re-election of Senator Harris hardly seems evidence that this was "the most important purpose of the Bourbons in the legislature" (p. 71). Finally, cooperation of Populists with Republicans seems less a sacrifice of programs for the purpose of "revenge" than a pragmatic pursuit of the power to affect policy. Redeemers, Bourbons & Populists provides a generally competent and occasionally provocative narrative of many political developments in Tennessee. Its failme may perhaps reflect a lack of attention to the growing literature on Gilded Age politics. These works point out the importance of a broader range of factors, like religion, and such social questions as liquor control, subjects which Hart virtually ignores. Closer attention to these works might also have minimized certain of the problems mentioned here, for while systematic analysis is difficult to achieve, other authors have been much more successful. NOTES 1. See Hackney's Popul&m to Progressivism in Alabama (Princeton: Princeton University Press, 1969). 2. "Post-Reconstruction Suffrage Restrictions in Tennessee: A New Look at the V.O. Key Thesis," Political Science Quarterly, 88 (December 1973), 655-83; and The Shaping of Southern Politics: Suffrage Restriction and the Establishment o f the One-Party South, 1880-1910 (New

Haven: Yale University Press, 1974). 3. See Hackney, Populism to Progressivism and Stanley B. Parsons, The Populist Context: Rural Versus Urban Power on a Great Plains Frontier (Westport, Conn.: Greenwood Press, 1973). Philip R. VanderMeer is an asistant professor o f History at Purdue University.

People-Oriented Computer Systems: The Computer in Crisis, Edward A. Tomeski and Harold Lazarus.

New York: Van Nostrand Reinhold Company, 1975. Pp. xx + 300. $13.95. By Richard A. Bassler Over the years technological innovations have stirred up emotional reactions on the part of the people affected by the resulting changes. In the early nineteenth century, workmen convinced that machines were the direct cause of their unemployment and low wages, systematically destroyed the offenders. Somewhat later, Karel ~apek's play R. U.R. prophesied the disastrous results of man's domination

374

REVIEWS

by machine. During the late 1960s and early 1970s the modern-day Luddites turned on the computer as the target of their frustration. An article on "How to Destroy a Computer" appeared in the Free Press in Los Angeles. As a response to this sort of antagonism, Tomeski and Lazarus have emphasized the humane use of computers in a book which recognizes people problems as the major barriers to the effective utilization of computers. According to the authors, their book will appeal to executives in various public and private organizations who have been frustrated by difficulties with computerization. The book consists of three major parts-"The Human System," "The Computer System," and "Computer Systems Serving Social Systems"augmented by a brief "Epilogue." Notable in the organization is the inclusion in each of the three main sections of illustrative case studies which help immensely in the reader's understanding of the complex relationships being developed. For classroom purposes, the useful discussion questions accompanying each case can be included in a seminar. Lest the reader infer, however, that this book is only for academic purposes, let me state that almost anyone involved with computers will benefit from the authors' thorough presentation of case studies. They have clearly drawn from their real-world experiences in developing their material and reaching conclusions. Their writing is clear and to the point, without the usual computer-world jargon understandable only to those familiar with "software." Important is the authors' concept of the relationships between the various elements in a working computer system. In the first part of the book, a base is developed that relates to human systems and identifies opportunities and challenges for the practitioner. The second part focuses on the computer to highlight the symbiosis among users of computers, fumishers of the computer media, and computer professionals attempting to put it all together. In the third part, the interthces between human systems and computer systems, which were discussed in the first two parts, are combined to serve the social system of an organization. In the Epilogue the authors suggest that we should no longer bend, staple, and mutilate people in achieving the goals of computer systems. The index, although adequate, could have been more inclusive. The bibliography accompanying each chapter is useful to the reader who wishes to pursue the subject further, and the footnoting is unobtrusive. This well written, well organized, and well conceived book should be read by all who purport to be

computer professionals. Soon the needs of the users and of the people involved in computer systems must be regarded as inseparable. These authors have indeed looked ahead to some of the approaching problems. What is more, they have offered some excellent advice for all of us. Richard A. Bassler is a professor of computer systems applications at the American University, Washington, D.C.

Les Banques de Donn~es Arch~ologiques: Colloques Nationaux du CN.R.S. no. 932. Marseilles, 12-14 juin 1972. Paris: Editions du C.N.R.S., 1974. Pp. 328.

Par Marie Christine Duchenne La cr6ation de banques de donn6es arch6ologiques soul,vent des probl~mes de m~thodologie, de technologie et des probl~mes institutionnels. Les problames m~thodologiques sont de trois ordres: le choix des crit~res 5 retenir pour individualiser un objet, les relations 5 6tablir entre les trois types de langage (le langage naturel, le langage scientifique et le langage documentaire) et enfin le probl~me du d&oupage pour d~terminer dans quelle banque sp~cifique de donn~es le document sera enregistr& Des probl~mes techniques interviennent au niveau des liaisons entre les diff6rentes banques de donn6es. Deux solutions sont envisag~es: l'unification des syst~mes d'exploitation ou le d6veloppement de banques de donn~es differemment con~ues, solution jug~e preferable ~tant donn6 le stade des "projets pflotes" des banques de donn6es arch6ologiques. Etant donn6 que la phipart des projets de banques de donn6es arch6ologiques ~manent d'initiatives personnelles, les probl6mes institutionnels comme ceux du financement des banques de donn6es et le leur entretien (corrections, int6grations) devraient amener les arcMologues ~ confier leur gestion et leur entretien un organisme patronn~ par des unions seientifiques internationnales, ce qui augmenterait le nombre des utilisateurs actuellement restreint. L'introduction de ces probl~mes s'accompagne d'une communication assez br~ve de Ph. Picard sur les apports des T616communications aux recherehes documentaires en France. La Pr~histoire fait l'objet de trois communications. La premiere, "Exploitation automatique des donn6es d'un site pr6historique", se compose de trois parties: H. de Lumley-Woodyear d6crit en particulier

REVIEWS l'~laboration des lexiques n~cessaires g l'enregistrement des informations et le traitement des informations pour la r6alisation de trac6s automatiques de projections verticales d'objets ou de plans. L. Bourrelly pr~sente le programme de recherche documentaire et Madame H. Camps-Fabrer un projet de code pour l'analyse d'objets en os. Ce projet se caract6rise par la cr6ation d'un lexique analytique de termes destin6s ~ l'6tude morphologique et technique de n'importe quel objet e n o s et par l'gtablissement d'une fiche descriptive utilisable au cours des fouilles et pour l'~tude d'objets e n o s conserv6s dans les mus~es. La communication suivante, "Contribution la cr6ation d'une banque de donn6es arch6ologiques", pr~sente bri6vement les probl~mes, les programmes et les premiers r6sultats d'un traitement sur ordinateur des donn~es arch6ologiques de deux gisements pr~historiques du Sud-Ouest de la France. D'autre part, G. Laplace explique les raisons pour lesquelles des crit~res morphotechniques et typom~triques scientifiques ont 6t~ 61aborfis pour fonder une typologie rationnelle avant de faire part de ses travaux sur le Pal~olithique. Plusieurs syst~mes d'exploitation automatique appliques au domaine des Collections G~n~rales sont pr~sent~s. E. Chouraqui, pour l'Inventaire G~n~ral des Monuments et Richesses Artistiques de la France, expose rapidement la m~thode d'analyse et d~finit le langage formel d'analyse retenu. La communication suivante sur "Le Syst~me d'exploitation automatique des Collections ethnographiques au Mus6e des Arts et Traditions Populaires" pose le probl~me en particulier des relations qu'il a fallu 6tablir entre plusieurs syst~mes documentaires. Les travaux r6alis~s 5 partir de la collection des objets domestiques fran~ais ont permis de mettre au point des techniques de description, ici d~velopp~es, pour la cr6ation d'une fiche portant des informations administratives et des informations ethnographiques. Par ailleurs, Madame A. Borillo expose la m~thodologie d'un syst~me int6gr~ de traitement de l'information textuelle et graphique sur la base d'un ~chantillon du Corpus des amphores antiques de la M~diterran~e Oceidentale; elle d~veloppe la question du traitement des questions en langage naturel et du traitement num6rique des donn~es graphiques. Dans le cadre de cette experience on envisage d'aider l'utilisateur par un lexique de traduction qui puisse tenir compte des ambiguit~s du langage naturel. Enfin les travaux r6alis6s pour l'exploitation automatique du Corpus des Inscriptions Latines, en rue de l'~dition du tome VIII du Corpus, d~finissent les outils documentaires: les codes formels et le langage documentaire.

375

La troisibme partie de cet ouvrage regroupe les exposds des travaux pr6paratoires du Centre de Recherche d'Arch~ologie Classique de Paris X pour la r~alisation d'une banque de donnbes couvrant plusieurs champs de l'arch~ologie classique. Ces travaux, conduits dans la perspective d'un accord avec d'autres systbmes plus vastes en vole de r~alisation-le Museum Computer Network et la transposition informatique sous la responsabilit~ du Centre d'Analyse documentaire pour l'ArcMologie de l'Inventaire G~n6ral des Monuments et Richesses Artistiques de la France-se proposent de cr6er un syst~me d'exploitation automatique fonctionnant non seulement en rue des operations de recherche documentaire mais aussi pour les applications de type math~matique. I1 s'agit ici en particulier des travaux effectu6s sur Les Mosaiques Grecques et Romaines pour la recherche d'un langage adapt~ d~s le d6part aux techniques d'exploitation de la banque de donn6es; des travaux effectu~s sur La Sculpture Romaine, domaine qui pose un certain nombre de probl~mes dus g la presence de deux types de donn~es: des donn~es objectives et des donn~es subjectives intervenant au niveau de la description des monuments. Les travaux r~alis6s sur Les St~les Fun6raires Hell6nistiques de Thessalie rendent compte des probl~mes pos6s par le choix des donn6es pour chaque st~le. Ces travaux font l'objet d'un expos~ sur le syst~me de description qui ne se fonde pas essentiellement sur les concepts des arcMologues mais sur les r~gles de fabrication des st~les. Enfin plusieurs projets concernant la documentation et la bibliographie arch~ologique sont pr6sent6s. J. Leclant pr6sente l'~tat des travaux sur l'Inventaire syst6matique de la documentation concernant la Civilisation M~ro~tique, travaux qui ont abouti l'enregistrement automatique de 900 textes m~ro'itiques. J. P. Demoule pr~sente le projet de Bibliographie Automatique en Pr6histoire et Protohistoire europ~enne, qui pr~voit le traitement des textes et documents iconographiques avec la collaboration de tousles pays concern6s. S. Cleuziou, dans son expos~ du projet de code pour l'analyse d'une collection de documents photographiques en arch~ologie, d~crit l'organisation du code 61abor~ en fonction des besoins des utilisateurs, pr6voyant six champs /i l'int~rieur desquels les termes du lexique sont hi~rarchis~s, et l'organisation du fichier. Par ailleurs, le projet d'exploitation automatique d'une Carte arch6ologique de la France nous est pr~sent~; il utilise le programme de l'Inventaire G~n6ral des Monuments et Richesses Artistiques de la France pour permettre un passage constant entre les deux banques de donn~es.

REVIEWS

376

Cette s~rie de communications s'ach~ve par une discussion ~t propos du d~veloppement actuel en France des r~saux de domin6es socio-~conomiques pour la pr6paration des d6cisions et par un aperqu des banques de donn6es arch6ologiques ~ l'6tranger.

Marie Christine Duchenne est membre du Centre de Documentation SOmitique, C.R.A.L., Universit~ de Nancy II.

A Concordance to Juan Ruiz "Libro de Buen Arnor', ed. Rigo Mignani, Mario A. Di Cesare and George F. Jones. Albany: SUNY Press, 1977. Pp. xiii + 328. $50.00. ISBN 0-87395-322-3. By Nancy Joe Dyer The goals of editors Mignani, De Cesare and Jones are "to work on the basis of an authentic a text as possible while endeavoring to make the concordance as useful as possible" (p. vii). The concordance, which took three years to prepare, was processed by an IBM 370/158 and provides a complete inventory of all words in the base text, including prose passages and high-frequency function words. The general technical "Introduction" gives the reader an orientation to manuscript and textual problems, to editorial principles, and to notes on use. Appendix 1 reproduces prose and fragments; Appendices 2 and 3 are word frequency lists, ordered alphabetically and by descending numerical frequency. The selection of a base text for the concordance was complicated by the lack of a definitive manuscript and by the number of critical editions which add to the sum of spurious material in the published "'traditional text." The editors' solution was to integrate three incomplete manuscripts (Salamanca = S, Toledo = T, Gayoso = G), none of which is the original, using S, the most nearly complete, as a base text "as if its forms must be closer to Juan Ruiz's original than the forms of G and T" (p. viii). The text was keypunched from the transcription of Criado de Val and Eric Naylor 1 then checked against the manuscripts themselves. Variants from G and T are limited to about 80 words which do not occur in the base text. Of the 210 cases of copyist error noted in the Criado de Val transcription, our editors correct about 50 obvious textual errors, but leave those forms which might have been dialectal. This editorial decision may provide future users with interesting clues about the development of the manuscript

tradition since although G and S are written in Castilian, T shows features of the Leonese dialect. In the entry hst, the editors admit to having "'normalized extensively, but not with absolute consistency nor out of pedantic compulsion. Our aim was to make the concordance more usable" (p. x). Showing sensitivity to the research needs of the philologist, they do not normalize possible dialectical peculiarities (l/r, y/ll, -t), although the nonsignificant spelling alternations are regularized. This reviewer suggests that all users, even those very familiar with graphemic alternations in Old Spanish manuscripts, refer to the rules used to "normalize" spelling, p. x. The Note to Users registers imperatives and infinitives with metathesized/assimilated enclitics (e.g., seguiIda = seguidla, tragalIos = tragarlos) which appear as individual listings in the entry list. Apocopared enclitics are included with the list of complete forms (e.g.,-m forms included with me). Researchers interested in grammatical or stylistic research should be aware that in the entry list the 350 homographs are separated by function (verbs and tenses, noun, adjective, pronoun). The homographs probably were not separated by the computer since, in the editors' words, "it remains to be seen whether the mighty labors necessary to produce these rarified results were worth the trouble" (p. xi). Page format and typefaces facilitate location of entries. The two columns of entries are separated by and flanked by generous margins. The page number is centered in the upper margin between the columns. Entries, set against the left margin in bold-face all caps, occasionally are followed by an editorial key, a letter in parentheses set in the same bold print. Most of the keys can be solved without explanation (H = homograph, L = Latin, V = variant). Indented, beneath the entry, the complete context or line of poetry is preceded by identification of the base manuscript, both stanza and verse. The context of prose in Appendix 1 is a meaningful syntactic group of adequate length, numbered sequentially P1 to P205 to facilitate inclusion in the concordance. Through thoughtful planning the editors have advanced the technology for preparation of computer-assisted research tools. Exemplary features are the paleographic base text, the aU-comprehensive entry list providing sufficient context and the inclusion of word-frequency lists, z Although minor editorial refinements could be desired, 3 Mignani, Di Cesare and Jones appear to have engineered a sophisticated and serviceable reference work. As with all such works, however, its ultimate usefulness will be adjudicated by scholars active in Libro de buen amor studies.

REV IEWS Notes I. Manuel Criado de Val and Eric W. Naylor, eds., Ltbro de buen amor, 2nd ed. Madrid: CSIC, 1972. The absence of a complete citation of this edition is a regrettable oversight, given Mignani's bibliographical expertise ("Bibliografia compendiaria sul 'LLbro de buen amor'," Cultura Neolattna, 25 [1965], 62-90), the concordance's open indebtedness to the pubhshed paleographic text and the presence of other btbliographical notes to the Introduction. To worsen the slight, Naylor (Eric Woodfin) is referred to as "Edward" (pp. vii, ix). 2. For a useful study of guidehnes for concordances in Old Spanish, see Steven 1). Kirby, "Concordances to Old Spanish Texts: Present Status and Proposed Future Guidelines," La coronica, 6 (1977), 38-40. 3. As previously indicated, the title, place and date of the pubhshed version of the base text, as well as the correct name of its editor, are not to be found in the concordance. The normalization of spelling for the entry list is inconsistent. In the Introduction, foreign words and parts of words, symbols and sigla are not distinguished by typeface or punctuation. Also, the reader would appreciate an explanation of abbreviations used in the concordance. Future editors attempting the division of homographic forms might wish to establish in the editorial norms a consistent sequence for listing homographic forms; for example, verb/noun, present/past. This reviewer hopes that these superficial flaws are not symptomatic of more serious, less immediately-apparent obstacles to the editors' goals.

Nancy Joe Dyer, an assistant professor of Spanish at Texas A&M University, & a specialist in Old Spanish language and literature.

377

vocabulary, unique words listed by manuscript, a complete inverse vocabulary, word frequencies for each manuscript, a list o f capitalized words, and a combined concordance o f all manuscripts. The format of Ruiziana, unlike that of its printed companion contributes nothing to its serviceableness. Microform is awkward to use for the necessarily frequent cross-references among the complementary constituent parts and between the microfiche and printed texts. Manageable hard copy reproductions are expensive and violate copyright laws. The typeset is computer print-out; the Preface offers no explanation for conversion o f manuscript graphemes to computer type. However, if the scholar can overcome the inconvenient medium of presentation, here is a splendid, innovative new tool for investigating linguistic, stylistic and textual aspects of an important literary work.

Nancy Joe Dyer, an assistant professor of Spanish at Texas A&M University, is a specialist #~ Medieval Spanish language and literature.

A Concordance to the Poetry o f Thomas Traherne. Ed. George R. Guffey. Berkeley and Los Angeles: University o f California Press, 1974. Pp. xv + 521. $30.00. ISBN 0-520-2449-4. By R. Lynn Sauls

Ruiziana, Research Materials for the Stud), o f the 'Libro de Buen Amor.' Ed. Rigo Mignani and Mario A. Di Cesare. SUNY Press, 1977. Microfiche. ISBN 0-87395-344.4. By Nancy Joe Dyer Demonstrating the vast usefulness of computerassisted data-retrieval techniques, editors Mignani and De Cesare have reprogrammed tapes prepared for their recent work A Concordance to Juan Ruiz 'Libro de Buen Amor'. The result, reduced to nine microfiche for economy, comprises an extensive collection of multipurpose research materials which should be of interest and use to scholars o f the Libro de Buen

Amor. Ruiziana features a combined line-by-line text of all manuscripts, an analysis of graphemic distribution, a list of text words ordered by length, a crossreference list of words with "normalized" spellings, and a list of words containing enclitics. Other potentially invaluable compilations are a comparative

A Concordance to the Poetry o f Thomas Traherne is a useful tool for those interested in Traherne's language and ideas, especially as they occur in his rhymed poems. It is based on the standard edition of Traherne's poems edited by H.M. Margoliouth (Centuries, Poems, and Thanksgivings, 2 vols. [Oxford, 1958]) and was computer programmed by Vinton A. Dearing. Professor Guffey's Preface concisely and clearly describes the scope of the concordance and explains the entries. Indexed are the rhymed poems that are extant in Traherne's hand (those from the Dobell Folio, "The Church's Year Book," "Philip Traherne's Notebook," and the "Centuries"), the poems in Philip Traherne's hand from the Burney Manuscript that are not extant in Traherne's hand, and the rhymed poems from printed books (those from

Christian Ethicks, Meditations on the Six Days of the O'eation, and The Thanksgivings). Entries in the concordance indicate the volume and page on which each word concorded occurs in the Margoliouth edition as well as the title and line

378

REVIEWS

number of the individual poem. Then the full line of poetry in which the word occurs is reproduced. Frequency of words, as well as relative frequency to the fourth decimal place, is given, not only for words indexed but also for the relatively insignificant words not concorded. As an example, the entry for "a," a word not concorded, reads as follows: frequency: 533

relativefrequency: 0.0152

The entry for "'eternity," a word that is fully concorded, begins thus: eternity frequency: 17 relativefrequency: 0.0004 See also "eternitie" II. 3 AUTHOR 60 Unto the Poles, and view Eternity. The "'II" refers to the second volume of the Margoliouth edition; "3" refers to the page; "'AUTHOR" refers to the poem, "The Author to the Critical Peruser"; "60" refers to the line of the poem. Appended to the concordance is a list of words in order of frequency beginning with the word "the," which occurs 1479 times, and ending with "'zodiac," the last of the words occurring only once. As a student of Traherne would expect, some of the words that appear most frequently in the poems other than structure words common in all English communication are "things," "soul," "God," "earth," "joy(s)," "world," "love" (also "lov"), "bliss," "appear," "glory" (also "glorious"), "divine," "sun," "skies," "shine," "treasure," "sight," "'gold," "eye" (also "ey"), "sense" (also "sence"), and "pleasure(s)." By comparing the concordance entries with cards I spent hundreds of hours preparing by hand when I analyzed Traherne's spelling preferences a few years ago, I determined that there is no reason to question seriously the accuracy of the completeness of the concordance, Such minor errors as the misspelling of "Zamzummim" as "Zamzummin" in the list of Words in Order of Frequency and the computer's inability to differentiate between Traherne's "then" which means "than" and his "then" that does mean "then" is inconsequential. A student of Traherne interested in the imagery or ideas in a given poem could quickly find other contexts for that imagery or for those ideas throughout Traherne's other rhymed poems. The concordance could save hours of searching. It is to be regretted, however, that the entire Thanksgivings and Centuries were not concorded along with the rhymed poems. The Thanksgivings were recognized to be a type of poetry, even in 1699 when they were first printed. And the Centuries,

which are generally considered to be Traherne's best work by far, are really prose poems. The language and ideas in one of Traherne's rhymed poems must be considered not only in the context of his other rhymed poems but also in the context of his best and most comprehensive works. A concordance of the poems including all the Thanksgivings and the Centuries would, of course, be much larger and much more expensive than the volume that Professors Guffey and Dearing prepared; but until such a concordance is available, or an index that could serve in its stead, the present work must be considered incomplete, though helpful, and therefore limited in its usefulness. If the expense of a concordance to the complete works of Traherne is prohibitive, then perhaps an index to the more significant words occurring in the major works, including Christian Ethieks and Meditations on the Six Days of the Creation, could be prepared. Guffey's

Concordance to the Poetry of Thomas Traherne would be indispensable to the preparation of such an index. R. Lynn Sauls is Professor of English at Atlantic Union College in South Lancaster, Massachusetts. Peiforming Arts Research: A Guide to Information Sources. Marion K. Whalon. Detroit: Gale Research Company, 1976. Pp. xi + 280. $18.00. ISBN 0-8103-1364-2. By R L Widmann This volume is an annotated, selected bibliography of reference tools, directories, periodicals, indexes, bibliographies, etc. For the most part, annotations are sharp, clear, and direct. The first part of seven divisions within the work is guides, subdivided into theater arts, dance, costume, visual arts and music, literature, rhetoric, general reference works, and periodicals. Part two is devoted to dictionaries, encyclopedias, and handbooks in fields such as theater, music, dance costume, and literature; part two concludes with a special section devoted solely to Shakespeare. Part three is a compilation of directories; part four is an incisive yet comprehensive recording of play indexes and finding lists. Part five offers sources for reviews of plays and motion pictures. Part six is devoted to bibliographies, indexes and abstracts for the theater and related arts, music in the theater, design in the theater, and motion pictures and broadcasting; this part also contains basic reference works in literature and in searching out dissertations and theses. Part seven is devoted to works on illustrative and audiovisual sources. The thirty-page

REVIEWS index is easy to read, even though it is printed in double columns. Whalon notes if a book, such as Marvin Spevack's multivolume Shakespeare concordance, has been produced with the aid of a computer. And there are, of course, omissions; Roger Manvell's important and basic book on Shakespeare and film is not mentioned, although Manvell is listed as general editor of the 1972 International Encyclopedia of Film. Nor is there mention of Robert C. Schweik and Dieter Riesner's Reference Sources in English and American Literature: A n Annotated Bibliography (U.S.A. publication, 1977; German publication, 1976) which supersedes the Altick and Wright Selective Bibliography for the Study o f English and Amel~can Literature (1967). Whalon does not always list the most recent editions of works; the Altick and Wright Bibliography has appeared in the fifth edition in 1975 and Altick's famous book, The Art o f Literary

379

Research, is listed as 1963, but has been revised and republished in 1975. In brief, this book is pleasant to handle and use. It has adequate white space in the margins for adding one's own annotations or listing call numbers for library copies of the volumes discussed. The audience for whom the work is designed would probably be advanced undergraduates and graduate students. Most scholars would already know the works surveyed by Whalon, but might like to have this handy volume in a personal library nevertheless, for it would be useful in reminding one of the range of basic works in a given field. The book is expensive at $18.00, but worth the money.

R L Widmann, former book review editor of CHum, teaches graduate courses in bibliography and literary methods at the University of Colorado, BouMer.

The Greek Text of the Gospels Before the Tenth Century is a project which has as its immediate purpose to compare the extant manuscripts of the Gospels in Greek dating from before 901 A.D., and to produce for each Gospel an apparatus criticus, a stemma, and an archetypal text. A secondary purpose is to illustrate some uses of computers in textual criticism. A longer-range purpose is to standardize the wording of the archetypal texts at places where they are synonymous and to produce therefrom an apparatus criticus, a stemma, and an archetypal Gospel, as a basis for a life of Jesus. The text-critical methods to be employed are explained in Vinton A. Dearing, Principles and Practice o f Textual Analysis (Berkeley: University of California Press, 1974). The computer programs to be used have been privately printed for copyright purposes and deposited in the Library of Congress and in the University Research Library at the University of California, Los Angeles. Prof. Dearing will be happy to discuss the use of computers with anyone who is interested in a similar project. The apparatus criticus for the major manuscripts of each Gospel will also be circulated in preliminary form, several chapters at a time. Comments on the significance of the variations recorded will be welcome, and will be circulated. Readers who would like to receive newsletters should send their names and addresses to Prof. Vinton A. Dearing, 2225 Rolfe Hall, University of California, Los Angeles, CA 90024.