The introductory overview of Harris’s oeuvre is by Harris himself, in a paper that has not previously been published in English. In the section on philosophy of science, Tom Ryckman discusses salient characteristics of Harris’s methodology of language and information. Paul Mattick, Jr. takes up some implications of Harris’s work for some questions in philosophy of science. Francis Lin argues that Chomsky’s well-known rejection of discovery procedures was mistaken.
Applications in informatics, information retrieval, automatic text generation, and translation, provide a kind of existence proof for Harrisian methodology and theory. These are discussed in the section on computational issues. The adjunction grammar of (1962) was the initial basis for Sager’s first demonstration of the computibility of language—not merely of the language-like symbol systems that are called, in what perhaps is no longer recognized to be metaphor, formal languages. The section on computability of language and sublanguage edited by Stephen Johnson surveys these developments and indicates current and future directions of research, with contributions by Naomi Sager, Ngo Thanh Nhan, Ralph Grishman, Daniel Gordon, Stephen Johnson, Richard Kittredge, Aravind Joshi, and Fernando Pereira. The paper by Harary and Helmreich lays out a mathematical basis in graph theory for new lines of research and application. It was important to Harris that his work should be socially useful, so it is entirely fitting that this section also shows some applications. Other contributions describe diverse applications to poetics (Kendall), pedagogy (Lukoff), and computer interface design (Smaby).
Harris (1946) introduced set-theoretic and algebraic notations as means for summarizing distributional facts in a form in which configurational relations and patterning are more easily discerned and manipulated than in the extensive tabulations of data that characterized Sapir’s way of working with languages. These conventions have since become inseparably a part of doing linguistics. In combination with systems of mathematical logic, and with the development of systems for programming computers, they have given rise to a proliferation of language-like systems and their grammars. Harris’s substitution grammar of (1946) was the starting point for Chomsky’s formulations of phrase structure grammar (PSG), whose pseudo-hierararchical structures of abstract nodes are a sine qua non of Generativist linguistics even today. Joshi exploited the complementarity of these two types of grammar in his tree-adjoining grammars (TAGs), supplementing the weakness of each by the strength of the other. Finally, the Operator Grammar of (1982) is close to dependency grammar and especially to categorial grammar, as well as bearing a superficial similarity to the predicate calculus of mathematical logic.
The formal analysis of language that Harris initiated has of course been famously developed along divergent lines by Noam Chomsky and his students. The notion of a generative grammar was sketched in Harris (1951[1947]:372):
The work of analysis leads right up to the statements which enable anyone to synthesize or predict utterances in the language. These statements form a deductive system with axiomatically defined initial elements and with theorems concerning the relations among them. The final theorems would indicate the structure of the utterances of the language in terms of the preceding parts of the system.
There may be various ways of presenting this system, which constitutes the description of the language structure. […] Other types of presentation which have frequently been used have depended ultimately on moving-parts models such as machines or historical sciences. In using such models, the linguistic presentation would speak, for example, of base forms […], of derived forms […], or processes which yield one form out of another. In all these types of presentation, the elements are seen as having histories, so that the relation of an element to sequences which contain it becomes the history of the element as it is subjected to various processes and extensions. […] In such presentations, a relation between two elements a and b is essentially the difference between two historical or otherwise derivational paths: that from A to a and that from A to b. A is set up as a base from which both a and b have, by different paths, been derived.
The fundamental difference between Harris’s work and that of Chomsky and his students is that Generative linguists assume the existence of a biologically inherited metalanguage antecedent to language, called Universal Grammar, whereas Harris assumes that there can be no prior metalanguage external to language itself. Chomsky and others interpreted this as Harris’s mere refusal to admit any reality to the structures identified in linguistics, but he only regarded that sort of commitment as premature. This is demonstrated by the far-ranging conclusions and conjectures found in Harris (1988, 1989, 1991). Paraphrasing Harris in another context, we may say that when we do not rest with the explanation that something is due to Universal Grammar, we may discover with additional hard work a formal regularity for which there can be a simpler explanation. Other differences between Harris’s work and Generativist linguistics follow from this contradiction in their underlying commitments. In particular, the ramification of complex abstract structure in a symbolic and graphical metalanguage contrasts with Harris’s use of the metalanguage that is inherent in language itself; the assumption of separate mechanisms for semantic interpretation contrasts with the disclosure of information structures in language itself; and the countless detailed studies of grammar fragments contrasts with complete coverage of the grammar from (1946) onward. It is interesting that the chief exception to this, Harris (1951a[1947]) was formative in Chomsky’s development.
In the section on syntax, Terry Langendoen presents developments in his own string theory for linguistic analysis, showing how it relates directly to string grammar as developed by Harris, Joshi, Sager, and others, and indirectly to Operator Grammar. Richard Oehrle investigates formal properties of operations that contatenate words by intercalating different strings, in relation to the linearization of an operator on its arguments in Operator Grammar. Haj Ross pursues a theme of indeterminacy (ill-defined or graded membership in grammatical catagories) that has occupied him since Harris alerted him to it some 35 years ago, this time looking at what he calls "decay" of some nouns from their nominal status. Pieter Seuren explores somewhat similar issues with constructions like "go fishing," "paint x green," and "put x up," which he terms pseudo-complements. Anne Daladier’s analysis of the passive construction in French proposes extensions to Operator Grammar, drawing upon the historical development of the passive in Indo-European languages. She emphasizes a Sapir-Harris perspective in contrast to the ontological commitment to linguistic universals that is more widely accepted in philosophy of language and in Generativist theory. Carlotta S. Smith investigates how co-occurance relations among features of a text indicate one or more points of view on its substantive report. Lila Gleitman shows how the strikingly rapid and robust acquisition of vocabulary by toddlers, despite the very great differences in the examples of usage that are available to them, is enabled by distributional information about verb subcategorization ("verbs of a feather"). Michael Gottfried extends work done in association with Harris (1989), showing that a sublanguage of science is closed under reference. Contributions by Benoît Habert and Pierre Zweigenbaum, Maurice Gross, and Morris Salkoff round out this section. [Abstracts not yet received.]
Introductory remarks in a number of papers place the work of Harris in historical context and provide historiographic information that has not previously been readily available.
Zellig Harris Home page | Book Description | References
This site is maintained by Stephen Johnson