Definition in Theory and Practice: Language, Lexicography and the Law

Citation

Bade, D. (2007), "Definition in Theory and Practice: Language, Lexicography and the Law", Journal of Documentation, Vol. 63 No. 6, pp. 987-992. https://doi.org/10.1108/00220410710836457

Despite the title, this book is not just for linguists, lexicographers and lawyers. Roy Harris has steamed up with University of Hong Kong Professor of English Christopher Hutton to produce a book that is as interesting and relevant to LIS as Harris' earlier books. Anyone interested in the problems of meaning and context in information retrieval and especially those focussing on the construction and use of thesauri, ontologies and the dream of the “semantic web” will find much thought for food here. The problem of reference (in the philosophical and lexicographical sense) is central to the authors' arguments and provides an unexpected illumination of problems of reference work in libraries as well. Rather than review the book in detail, my intention is to discuss certain parts of the authors' methodology and arguments as these relate to various aspects of meaning in current discussions within LIS.

The most significant barrier to the appropriation and use of the literature and theory of linguistics in librarianship and information science is the former's long‐standing rejection of communication as the foundation of language, the claim that communication is irrelevant to the study of language. From Saussure to Pinker the language of “linguistics proper” has not been Saussure's “parole” – the empirical facts of the actual linguistic activity of human beings – but his “langue” – the linguist's model of the system underlying “parole”. Lyons (1977) specifically noted this disjunction as something non‐linguists do not realize:

It is not generally realized by non‐linguists how indirect is the relationship between observed (or observable) utterances and the set of grammatical sentences postulated (and cited by way of example) by the linguist in his description of any particular language…

What the linguist does when he describes a language, English for example, is to construct what is commonly referred to by scientists as a model, not of actual language‐behaviour, but of the regularities manifest in that behaviour (more precisely of that part of language‐behaviour which the linguist defines, by methodological decision, to fall within the scope of linguistics): he constructs a model of the language‐system (Lyons, 1977, Vol. 1, pp. 26, 29).

Since librarians, library users and the information technologies which they use have to deal entirely and only with actual usage and that “actual usage in its totality is chaos” (Definition in Theory and Practice, p. 79), the librarian's and information system designer's task is exactly like the task of the lexicographer as Harris and Hutton describe it: “to reduce chaos to order”. No lexicographer, they insist, has “ever undertook to document chaos”; rather the task of the lexicographer, like the librarian, is to bring some order to that chaos, to engage in interpretation and sense‐making.

The authors of Definition in Theory and Practice (hereafter, DTP) are not interested in constructing a model of any in‐your‐head language‐system. They assume the basic tenets of integrational linguistics, including the assumption that linguistic signs are not the prerequisite for communication but are instead the products of communicational activity, this activity requiring the making of signs. One of the consequences of that theoretical presupposition is that language must be studied within the context of that communicative activity. How that presupposition affects linguistic research and theorizing can be demonstrated by contrasting a remark by Chomsky (1957) in Syntactic Structures with remarks in the book under review:

Syntax is the study of the principles and processes by which sentences are constructed in particular languages. Syntactic investigation of a given language has as its goal the construction of a grammar that can be viewed as a device of some sort for producing the sentences of the language under analysis. More generally, linguists must be concerned with the problem of determining the fundamental underlying properties of successful grammars. The ultimate outcome of these investigations should be a theory of linguistic structure in which the descriptive devices utilized in particular grammars are presented and studied abstractly, with no specific reference to particular languages (Chomsky, 1957, p. 11).

[T]here is no more need for postulating invariant collective codes that allegedly make it possible for one individual to exchange thoughts with another… Communication in all its forms is permanently at the mercy of circumstances and human reactions that are, in the final analysis, unpredictable. Which amounts to saying that the serious student of human communication cannot expect that study to be a “science” (DTP, p. 69).

We can profitably compare the differences between these approaches with the differences in LIS between Shannon and Weaver's mathematical approach and Brenda Dervin's “Communication, not information” approach. We know that much has been accomplished along the lines of Shannon and Weaver's mathematical theory of communication, and the usefulness of current information systems are to a large extent testimony to the value of the mathematical theory for developing communication technologies. Yet Dervin's work (among others) suggests that the theoretical assumptions of the mathematical theory and the interpretations which they impose do not do justice to the actual facts of human communicative behaviour in spite of their usefullness. If we wish to understand actual human communicative activity in the context of information production, seeking and use, we must seek a different theory.

The search for an abstract system in linguistics is pursued with a reliance on “intuitions” of grammaticalness and decontextualized bits of text that Lyons referrred to as “system‐sentences”:

[S]ystem‐sentences are sequences of words in a one‐to‐one order‐preserving correspondence with what would be judged, intuitively by native speakers, to be grammatically complete text‐sentences…

[S]ystem‐sentences never occur as the products of ordinary language‐behaviour. Representations of system‐sentences may of course be used in metalinguistic discussions of the structure and functions of language; and it is such representations that are customarily cited in grammatical descriptions of particular languages (Lyons, 1977, pp. 30‐1).

This issue is at the heart of the difference between Harris and Hutton's approach in DTP on the one hand and theoretical linguistics not oriented toward the integrational approach on the other. Harris and Hutton look at definition not only as theorized by philosophers, linguists, lexicographers (Hutton himself having published a dictionary of Hong Kong Chinese slang) and lawyers, but also as practiced, examining the definitions found in writings of all of the above, in dictionaries, court cases and legislation. Again, a comparison with Chomsky is illuminating. Four pages after the passage from Syntactic Structures quoted above, Chomsky's first and most famous example appears as a demonstration that grammatical sentences of English need not be meaningful:

Second, the notion “grammatical” cannot be identified with “meaningful” or “significant” in any semantic sense. Sentences (1) and (2) are equally nonsensical, but any speaker of English will recognize that only the former is grammatical.(1) Colorless green ideas sleep furiously.(2) Furiously sleep ideas green colorless (Chomsky, 1957, p. 15).

With these examples, we are a long way from English, and even further from communication.

In contrast, Harris and Hutton offer no hypothetical definitions, no imaginary scenarios, no abstract systems of defining, and no nonsense. What they offer instead is a careful presentation and critical analysis of four types of theories of definition as these are found in the literature on definition (stipulative definition, common usage, real definition and ostensive definition) on pp. 1‐74, followed by an examination of various lexicographers' understandings of their task and their practices as they are evident in their dictionaries (pp. 75‐130) and a discussion of how words are defined, contested, argued and redefined in the practice of law (including the use of “authoritative” dictionaries in court), basing their discussion on about 50 actual cases from Partridge v. Strange & Coker (1553) to an uncited case reported in the Daily Telegraph in 2005 (pp. 131‐95). In Part Four, the Conclusion, the authors discuss indeterminacy, reference and deconstruction in linguistics, philosophy, literary theory and law, distinguishing their position from that of deconstructionists, Legal realism and Critical Legal Studies. Stanley Fish gets a careful critical reading (“his more general argument about the nature of texts and interpretative communities seems to rely on the very distinction between ‘inside’ and ‘outside’ that he rejects in law” (DTP, p. 219), while Popper's belief that “we never know what we are talking about” is dealt with briefly: “For one who never knew what he was talking about, Popper talked a great deal, and convinced many naive scientists to the contrary“ (DTP, p. 222).

The authors' decision to study the theory and practice of definition in the construction and interpretation of law was due to their recognition that while all of what they call the “supercategories” of our time (such as History, Science and Art):

… are, at least to some extent, linguistic constructs,… the law is unique among them in the extent to which it relies overtly upon the possibility of determining verbal meanings. Appeal to the dictionary has become one of the features of contemporary jurisprudence. This presupposes the possibility of integrating lexicographical practice with legal practice. Second, the law is in any case a practical, institutionalized attempt to implement one particular type of integrational procedure; namely, integrating (1) the past verbal activities of legislators and testators with (2) the present and future activities (verbal or non‐verbal) of all those affected or potentially affected by (1). The form such integration will take, or should take, it is the function of judges and courts to decide. Without that integrational function, the law would have no raison d'être (DTP, p. viii).

By describing the function of the law as the integration of past, present and future activities relying upon the creation and integration of linguistic signs, the authors have sought to understand communication in law in the same manner as Hjørland has insisted that librarians must approach communication in the world of information and information users. Without the need to integrate the world of information with the present, future, and potential users of that information, LIS would also have no raison d'être.

The integration of information, information users, information professionals, information technologies and social agent technologies on the web is itself an interesting enough research area begging to be approached from the perspective of integrational linguistics. Where it gets really interesting – and where this particular book has so much to offer – is in the context of the conceptualization, construction and perhaps realization of the Semantic Web, a vision first articulated by Berners‐Lee et al. (2001) in their article in Scientific American. Since this Semantic Web relies primarily on a system of definitions and interpretive rules (called “ontologies”) to make the meaning of everything on the web (!) precise (!!) and unambiguous (!!!), one might expect to find amongst the vast literature on ontologies and the Semantic Web some careful discussion of definition informed by the philosophical, linguistic and lexicographical literature. Yet Alas! and Alack! Thou mayest Seek, but Thou shalt not find. All the more reason, then, for researchers in LIS to welcome Harris and Hutton's contribution to the conceptualization and articulation of the Semantic Web (even though they never mention it).

The easiest manner of demonstrating the pertinence of DTP to some of the most current and urgent issues in LIS – like controlled vocabularies, ontologies and the Semantic Web – is simply to juxtapose passages from that book with passages from recent articles like the Berners‐Lee piece mentioned in the preceding paragraph. Here we go:

The Semantic Web is not a separate Web but an extension of the current one, in which information is given well‐defined meaning … (Berners‐Lee et al., 2001).

[A]ll signs (not only linguistic signs) are semantically indeterminate. In this perspective, semantics is the study and practice of human attempts to impose some degree of communicational determinacy on signs. The successes, failures and limitations of such efforts are, in our view, central to the enterprise of definition (DTP, p. viii).

Ontologies furnish the vocabulary necessary for communication between social agents and Web pages, defining relations between concepts… in practice, an ontology defines terms associated with the texts it describes, what they mean themselves, and the formal axioms that restrict the interpretation and use of those terms (Pickler, 2007, p. 72).

A definition can only be as effective as the context allows it to be, and the context includes the situation of the person seeking to understand the meaning. The notion of a definition adequate to all occasions and all demands is a semantic ignis fatuus (DTP, p. 49).

An ontology defines the classes and concepts of an area of knowledge or of a process, offering the definition, the characteristics and attributes, and the relations that are established as well as the functions that complete the elements. Ontologies codify the knowledge of a domain and are made to be reusable (Romaní, 2006, p. 9).

The problem for the general lexicographer, who purports to be giving definitions for the total inventory of words in languages like English and French, is that the technique so successfully deployed in restricted domains such as physics and chemistry cannot be extended to general vocabulary … The reason is that words have no general referential focus, nor even anything that approximates to one: the very idea is self‐contradictory. Focus implies the deliberate exclusion, from a particular communicational context or type of context, of all facts deemed irrelevant thereto … That is why those who compile general dictionaries are forced to assume, whether they like it or not, that certain facts about the world are of more interest than others … Their definitions are tailored to those assumptions of referential focus, whether they realize that consciously or not (DTP, pp. 212‐13).

A program that wants to compare or combine information across the two databases has to know that these two terms are being used to mean the same thing. Ideally, the program must have a way to discover such common meanings for whatever databases it encounters (Berners‐Lee et al., 2001).

The postulated semantic consensus underlying everyday talk about cats and dogs is a theoretical illusion conjured up by generalizing across separate, individual decisions to call Fido a dog and Tibbles a cat for the communicational purposes to hand. But it is an idle cog in the explanatory wheel. For even if it were true that all members of a linguistic community had reached an agreement in advance on what to call Fido and Tibbles, their implementation of that agreement would still require an individual decision by each of them on each occasion (DTP, p. 202).

Ontologies, as “explicit representations of shared understanding” can also be used to codify the terminology's semantics. For example, it must be assumed in using XML that the author and reader of < foo > 7</foo> have the same understanding of what “foo” means (Kim, 2002, p. 52).

[N]o form of definition can render the meaning of a sign determinate, i.e. can guarantee that those who have either formulated or understood the definition have thereby grasped the same meaning. Definition, in short, always allows room for semantic divergence and hence misunderstanding. It may, depending on the circumstances, reduce uncertainty but cannot eliminate it (DTP, p. 65).

At the heart of the semantic web, ontologies are used to add a semantic layer to the actual web. They are references for communication between machines but also between machines and humans by defining consensually the meaning of objects firstly through symbols (words or phrases) that designate and characterize them, and then through a structured or formal representation of their role in the domain (Hernandez et al., 2007, p. 144).

[T]here is no particular underlying consensus about the meanings of words, because the routines of everyday interaction do not require it (DTP, p. 173).

An ontology is not a taxonomy, a classification scheme or a dictionary. It is, in fact, a unique representational system that integrates within a single structure the characteristics of more traditional approaches such as nested hierarchies, faceted thesauri and controlled vocabularies. An ontology provides the semantic bases for metadata schemes and facilitates communication among systems and agents by enforcing a standardized conceptual model for a community of users (Jacob, 2003).

All definitions purporting to capture orthodox or “good” linguistic practice are inherently stipulative: in effect, they instruct the reader “This is how the word should be used and understood”. They are stipulative in just the same respect as the Highway Code is stipulative. It tells one how people should behave on the road, not how they actually do behave. It proposes a model to be followed, just as a dictionary does … Linguistic “standardization” is an inherently prescriptive notion (DTP, p. 91‐2).

If we are going to have programs that understand language, we will have to encode what words mean. Since words refer to the world, their definitions will have to be in terms of some underlying theory of the world. We will therefore have to construct that theory, and do so in a way that reflects the ontology that is implicit in natural language (Hobbs, 1995, p. 819).

An integrationist theory of reference … has no room for the idea of static relations between words and the world (DTP, p. 208).

A linguistically‐based ontology corresponds to the way people think about objects. It is a useful way to predict their thinking about the knowledge in structured databases (Dahlgren, 1995, p. 810).

Thus the postulate of semantic determinacy makes it possible to transfer the responsibility for making meaning from the individual to the collectivity, and from the circumstantial to the macrosocial level … Words come to be construed as autonomous signs that can and do function semantically without support from any other semiological source, providing the essential self‐sustaining mechanisms for the facilitation and regulation not only of human interactions but of human thought itself (DTP, pp. 203‐4).

Librarians and other users and designers of information technologies should be especially challenged by that last quotation for at least two reasons. If indeed words do not regulate “human thought itself” then we can expect a measure of indeterminacy of meaning and mismatches between the language of the user and both the information we deal with and the metadata associated with that information. Furthermore, although the metadata we create cannot be understood to be “enforcing” interpretations on the library users/information searchers, the manner and context of its creation will significantly determine its communicational value. Since the objects with which we deal – printed texts, recorded sound, visual materials, digital objects of all sorts – must all be interpreted, the first question is who will do that interpretation for its present, future and potential users and uses? The author/creator of that information? An alive and engaged human being with a stake in the consequences of interpretation? An information professional with an eye on profit? An information proletariat counting the piece‐work or watching the clock? Software written by a programmer who believed that disagreement, debate, disinformation and misunderstanding do not exist or are irrelevant? Using their examination of the practice of definition in lexicography and law Harris and Hutton have argued and demonstrated that “signs do not ‘have meanings’ but are ‘made to mean’” (DTP, p. 223). This, they conclude, “points to the only basis on which human communication can – and does – proceed” (DTP, p. 223). And that points us to the basis for a theory of communication for LIS.

Read this book and rejoice.

References

Berners‐Lee, T., Hendler, J. and Lassila, O. ( 2001 ), “The semantic web ”, Scientific American , Vol. 284 No. 5 , pp. 34‐43 .

Chomsky, N. ( 1957 ), Syntactic Structures , Mouton , The Hague .

Dahlgren, K. ( 1995 ), “A linguistic ontology ”, International Journal of Human‐Computer Studies , Vol. 43 , pp. 809‐18 .

Hernandez, N., Mothe, J., Chrisment, C. and Egret, D. ( 2007 ), “Modeling context through domain ontologies ”, Information Retrieval , Vol. 10 , pp. 143‐72 .

Hobbs, J.R. ( 1995 ), “Sketch of an ontology underlying the way we talk about the world”, International Journal of Human‐Computer Studies , Vol. 43 , pp. 819‐30 .

Jacob, E.K. ( 2003 ), “Ontologies and the semantic web ”, Bulletin of the American Society for Information Science and Technology , Vol. 29 No. 4 .

Kim, H. ( 2002 ), “Predicting how ontologies for the semantic web will evolve ”, Communications of the ACM , Vol. 45 No. 2 , pp. 48‐54 .

Lyons, J. ( 1977 ), Semantics , Cambridge University Press , Cambridge .

Pickler, M.E.V. ( 2007 ), “Web Semântica: ontologias como ferramentas de representação do conhecimento ”, Perspectivas em Ciência da Informação , Vol. 12 No. 1 , pp. 65‐83 .

Romaní, M. ( 2006 ), “Webs semàntiques, les webs de segona generació ”, Item: Revista de Biblioteconomia i Documentació , Vol. 42 , January‐April , pp. 7‐19 .