published, but as this is a foundations or historical essay, it seems appropriate to include it in a set of Workshop papers which seeks to define the domain. Hudson tries to define ''main-This book is a collection of papers from a Workshop held stream'' linguistics and ''folk'' linguistics but more successin 1986 in Marina di Grosetto, Italy. The introduction written fully defines and elaborates upon the concept of a lexicon, lexiby (the late) Donald E. Walker et al. defines the purpose of the cal entry, and the like. Much of the rest of this foundations publication as putting the Workshop in a historical perspective paper is spent defining ''word grammar'' especially as it relates and confirming that the overlying questions have neither to the ''mainstream.'' This paper is strengthened by an appendix changed nor been adequately answered in the past decade.
which describes more recent developments in lexicography. While this collection would have benefitted from an updated Since this is the foundational paper, working definitions seem list of collective references, much of this material stands up appropriate here. well 10 years later. As is sometimes the case with collections B. Levin's paper, ''Approaches to Lexical Semantic Repreof papers, the indexing is rather shallow and is heavily comsentation'' could more appropriately be labeled lexical semantic prised of proper names; following the conceptual threads representation of verbs. Verbs arguably can be more difficult through the Workshop within the index would have strengthto assign meaning to than nouns, and more specifically, Levin ened the potential usefulness of this collection as a resource.
tackles the assignment of meaning to ''verbs whose arguments This collection could also benefit from a cohesiveness check: are expressed as noun phrase and prepositional phrase comple-Because each paper can and does stand alone, there is much ments'' (footnote to paper). After critically assimilating the repetition of fundamental and definitional points.
background of strengths and weaknesses of various ways of It is difficult to analyze a collection as a whole on the points attributing meaning to verbs along with their syntactic properenumerated in a traditional book review such as style, cohesiveties, Levin evaluates two of the more widely used lexical semanness, and theme, but it should still be possible to assess whether tic representations: Predicate decomposition and semantic role the stated purpose of publishing this collection a decade after lists. She makes the popular argument that the meaning of a the Workshop: That ''the Workshop . . . marked a turning verb is in a co-dependent relationship with its syntactic properpoint in the field . . . [was] a point of departure . . . addresses ties. This paper succeeds in providing a baseline through a well many issues still being debated today and that guide research organized review of approaches. and development efforts'' can be addressed. This cannot be R. J. P. Ingra's paper, ''Lexical Information for Parsing Sysdone without caveat, since some of the papers from the Worktems: Points of Convergence and Divergence'' begins with a shop have already been published and therefore were not inpractical overview. Ingra discusses the lexical information cluded in this collection. It is, however, possible to conclude which needs to be supplied to a parser; and how that information that this collection is a valuable contribution to the field for has generally been represented in lexicons. After supplying reasons which will be enumerated below. First, it will be useful practical information and examples, Ingra discusses how lexical to look briefly at key papers and their central points.
information might be shared among different systems. This is The ''Introduction,'' by D. E. Walker, A. Zampolli, and N. still a conundrum, but Ingra gives several examples of projects Calzolari, describes the significance of the Workshop and the which have attempted to do just that. While this paper contains motivation behind the publication of this collection. The intervaluable information and an updated list of references, it is disciplinary, multilingual participants of the Workshop arrived almost too densely populated by figures and tables and brief at a ''baseline for current state of research and practice in the program descriptions to be readable. field at that time'' and a list of recommendations. In some ways this paper is the most pivotal and timely. By synthesizing the ''The Lexicon in Text Generation: Progress and Prospects,'' work of the Workshop participants into a list of 33 recommendaby S. Cumming is based upon the Workshop, but it has gone tions, and their subsequent developments, it is clear that it will through several iterations, according to its footnote, and this take much more than a decade to resolve these issues which refinement is apparent. Again, a strong background for the text will modify and grow as the technology to automate the lexicon generation lexicon is laid. Differences between human-and continues to develop. To reinforce this, developments in the machine-based lexicons are defined; directionality in linguistic field and especially in automated lexicon building, which could processing is discussed, and updated references have been be directly or indirectly traced to the Workshop, are described.
added. The thesis is that the Marina di Grosetto Workshop created the ''The Role of Dictionaries and Machine-Readable Lexicons right collegial atmosphere to spur researchers and developers in Translation'' by J. Slocum and M. G. Morgan covers the on at a more rapid pace. Whether or not this is true, the introducmultilingual aspects of automating the lexicon and makes a tion makes a strong case.
case for the dual functionality of dictionaries, which is to say dictionaries readable by both humans and machines. Unfortunately, some of the material about public domain dictionaries