Definiteness

Cambridge Textbooks in Linguistics

by Christopher Lyons

Definiteness is the concept expressed by definite articles such as "the", and also by other expressions such as demonstratives ("this", "that") and personal pronouns. This textbook surveys such expressions in the many languages of the world, and also examines the theoretical literature on this aspect of grammar, to establish what definiteness is and how languages can express it.

Genres: Linguistics, Nonfiction

404 pages, Paperback

First published January 1, 1999

Reviews

[Setayesh Dashti · Rating 4 out of 5]

I think this is a solid introduction to the subject. It is of course a little bit outdated; the more recent works on maximality, for instance, are not included. But I cannot think of a more recent book that covers the broad concept of (in)definiteness both cross-linguistically and theoretically. The book is a little bit too syntactically-anchored and the semantic discussions, especially the formal semantic ones, are too concise. The syntactic arguments for a Definiteness functional projection seems interesting to me (a non-expert in transformational syntax), but I doubt it has been taken seriously, which would be a shame. The cross-linguistic data is nonetheless praiseworthy.

Also, I find it quite interesting that there is almost no trace of Christopher Lyons on the internet. Recent review articles on definiteness do not even cite this book. Another dirty secret of the field, perhaps?

[Gary Bruff · Rating 4 out of 5]

I would encourage all who are interested in typological variation and other challenges for Universal Grammar to delve into Christopher Lyons’ big book of THE. If you do, you will likely be surprised by how incredibly well the use of this little English word manages to perform much of the heavy lifting in contextualizing everyday speech and writing. Typically, the definite article and other determiners serve to indicate the speaker’s belief that a certain, often implicit, nominal reference is cognitively accessible to the hearer. Determiners like THE are grammaticalized differently in different tongues and might not surface (exist?) in a particular language at all. Syntactically, definites are typically determiner phrase projections, functional material that provides a scaffold for merging nominals with higher predicates. Semantically, definiteness can often lay down an anaphoric avenue for logical interpretation while also contextualizing sense and reference according to a language’s information structure routines. And pragmatically, definiteness indexes some aspect of the speech situation itself, pointing to a nominal’s reference while simultaneously shifting in a dynamic way the place of interlocutors in their social universe. Analysis in terms of these three sub-fields of grammatical knowledge yields some evidence that definiteness expressions emerge and are grammaticalized according to the demands of natural language rhetoric. Or of rhetorics in the plural, for cross-linguistic research has ascertained that determiners are neither universal in their structure nor uniform in their application.

In English at least, definiteness and indefiniteness appear to be simple enough. Definiteness in English may be marked not only via the definite article ‘the’, but also with demonstratives like ‘this’ and ‘that’, pronouns like ‘they’ and ‘them’, or possessives like ‘their’. (Lyons notes that English syntax is unusual in slotting possessives as definite determiners.) All of these markers of definiteness serve to indicate that the speaker’s intended reference is accessible to the hearer, whether because of prior mention, inclusiveness, uniqueness, or ostension. Superlatives and proper names also signify unique entities, and so they are definite as well. On the other hand, so-called indefinites really signal cardinality, indicating a number or amount. Singular cardinality articles typically derive from, or are identical to, a language’s word for ‘one’. For example, the English ‘an’ is cognate with ‘one’. The plural of a/an is sm, a cardinal term which is a reduction of the pro-form ‘some’, as in ‘she took sm time getting here.’ Importantly, cardinality does not necessarily entail indefiniteness, so ‘a/an’ and ‘the’ seem not to form a consistent and unified grammatical category (‘articles’), in English or elsewhere.

It is useful to think about definiteness in terms of the syntactic category of D, found mostly with nouns but often interacting with other lexical and functional categories. Apart from articles and pronouns, definiteness might surface in languages variously as definite adjectives, as subject and object agreement markers on verbs, or as separate definite conjugations on verbs. Some languages allow definiteness morphemes to be stacked in the syntax, licensing the equivalent of ‘the my house’ or ‘the Socrates.’ Lyons builds on Postal’s “On the So-called ‘Pronouns’ in English” in repeatedly calling attention to the unbroken links, both synchronic and diachronic, between pronouns and articles. Whether an NP projects a DP generally depends on the valuing of some definiteness feature on noun phrases. It is possible to have an NP without a DP, as in the case of indefinite noun phrases. But it is also possible to have a DP without a (surface) NP, as in the case of pronominal forms. Because we can say ‘we anthropologists,’ it follows that pronouns should be thought of as determiners. Hence, a DP occurring alone can be regarded as an article whose NP complement has been deleted, much as in elliptical expressions. When viewed across the centuries, many languages display a similar grammatical evolution or parallel drift, whereby demonstratives over time become functionally extended to the categories of personal pronouns and definite articles. This happened with Latin/Romance. Latin demonstratives like ille and illa became in French the pronouns il and elle as well as the definite articles le and la (which happen to have a broader distribution than English THE). The same thing occurred in Germanic development. Thus, those English words starting with /ð/ (this, that, these, those, they, them, their, thee, thine, thou, the) are all definite determiners derived historically from demonstrative morphemes in old Germanic. In both the Germanic and Romance cases, a grammaticalization of indentifiability reconfigured an older indication of local deixis into a new signification that a referent is cognitively accessible to (and so identifiable by) the addressee.

Semanticists like Barbara Partee have pointed out how definiteness expressions tie together old and new information, marking old information as definite because it is recoverable from the text or context. In “Nominal and Temporal Anaphora,” Partee explains how tenses are used anaphorically, much like definite articles and pronouns. Past tense, for example, is anchored to the immediate speech context in shifting a cognitive cursor to a state or event prior to the speech event itself. Tenses do not refer to an antecedent stricto sensu the way articles or pronouns do, but they are context dependent as much as articles and pronouns. And adverbs like ‘whenever’ can universally quantify over events (‘whenever he does X’) much as nouns can be universally quantified by ‘all.’ Partee also notes the type shifting inherent in the semantics of nominal anaphora. The usage of THE has the power to shift a nominal’s reference mapping from a function over subsets of a type to a function of individuation of a specified token or tokens.

Social distance or proximity are often marked by prototypically-definite indexical expressions as part of their polysemic functioning. In his perspicacious “Shifters, Grammatical Categories, and Cultural Description,” Michael Silverstein enumerates the many things speech events can do besides referring to objects or events in the semantic sense. Speakers do refer, but this is just one aspect of ‘meaning.’ Leaving writing aside, an indexical might hide or highlight relations between hearer and speaker, possibly by indicating the sex of the speaker or hearer as in some Native American languages, or by marking deferential relations as in Korean and Japanese, or else by marking the observance of a taboo as in the mother-in-law vocabulary in some Australian languages.

Chase Raymond’s study called “Linguistic Reference in the Negotiation of Identity and Action” examines how second person indexicals (words meaning ‘you’, whose meaning emerges by indexing the context of the speech event itself) are shifted in Spanish oral narrative to indicate a number of social and cultural distinctions. The decision of when to use tu and when to use usted (when speaking to a singular hearer) depends on both the pre-established statuses of the hearer and speaker, as well as on the improvised stances that the speaker and hearer might be taking in the discourse. Raymond looks at a number of transcribed pieces of talk which show how social meaning is constructed on the fly. Thus, a 911 call begins with the caller using ‘tu,’ common in customer service calls, but then shifting to ‘usted’ to mark an awareness that the current talk is institutional and somewhat formal and legalistic. In another example, a mother and her grown daughter address each other as ‘tu’ in a familiar way, but then the mother switches to ‘usted’ when chiding her daughter about something before switching back to ‘tu.’ Social distance and social proximity are thereby reinforced, over and above the purely referential and definite uses of pronouns. As did Goffman and Silverstein, Raymond understands the pragmatic role of pronoun choice as an “emergent and co-constructed feature of the ongoing discourse.”

Lyons makes two generalizations which I believe to be typologically significant. One is that languages generally will have either a topic marker or a definite article, possibly neither but certainly not both. This implies that topics are definite and that definites are topical. The second is his observation that languages with definite articles are most frequent in Europe and around the Mediterranean. In other words, unrelated languages like French, Hebrew, Hungarian, and Turkish have definite determiners, projected as syntactic DP nodes, while languages from other regions frequently do not.

So what do languages do without a THE? I have studied two such languages, Latin (whose progeny surprisingly all have articles) and Mandarin (which syntactically specifies a topic). Beginning with Mandarin, Peter Jenks, in his “Articulated Definiteness without Articles,” claims to have found an anaphoric/unique contrast in Mandarin grammar. Contextually unique entities (like “the sun” or “the Alps” in English) are marked with bare nouns in Mandarin, while previously mentioned entities (anaphoric) are marked with a demonstrative. Matrix subjects are topics in Mandarin, and so function as definite regardless of whether or not a demonstrative is used there. Jenks finds this contrast between uniqueness definites and anaphoric definites in Mandarin to be the same distinction as is marked in German preposition phrases, which have strong/anaphoric (‘von den’) and contracted/uniqueness (‘vom’) forms. Jenks maintains that this parallel between Mandarin and German justifies positing a parameter setting in Universal Grammar which can be evaluated in the same manner by historically unrelated languages.

But these findings are refuted in David Bremmers et al., “Translation Mining: Definiteness across Languages (A Reply to Jenks)”. Translation mining is a methodology whereby books which has been translated into multiple languages, in this case Harry Potter books translated from English into German and Mandarin, are analyzed to determine how a grammatical structure in one language corresponds with the structure in another language found in the same semantic context. What the authors found was that in the semantic field of definiteness marking, Mandarin and German are nothing like each other. One possible explanation for the discrepancy could be found in Jenks’ use of Taiwanese informants, as opposed to Bremmers’ use of a translation emanating from Beijing. While both Taiwanese Guoyu and Beijing Hanyu are called ‘Mandarin’ by speakers of English, it is hard to see them as the same language these days. Calling Taiwanese Guoyu ‘Mandarin’ is more of a political concession than a linguistic reality. Also, the oral/written contrast between the two data sets is sure to be significant as well.

Moving on to article-less Latin, David Wigtil’s “Latin Definiteness and English Articles” in a very different way shows how translation just may provide the key to understanding English definiteness. Using notional criteria, and without the hindrance of donkey sentences, Wigtil specifies the knowledge of English grammar required in order to furnish an English translation of Latin with articles. The translator must judge from the sense of the Latin text when a reference is ‘presumed knowledge,’ part of ‘the common world of reference,’ or else indicative of ‘one-of-a-kind elements.’ Moreover, while Latin possessives and demonstratives can mark a candidate for definiteness, definite nominals are often realized as zero in Latin. In other words, it is common in Latin for the definite reference to go unexpressed, since pronouns (especially subjects or objects of verbs), along with silent entities modified by adjectives (‘stultus’ meaning ‘the fool’ instead of ‘foolish’) or by demonstratives (‘ille’ meaning ‘that one’ or just plain ‘he’ instead of ‘that’) often go unexpressed precisely because they are identifiable from the context. Wigtil looks at indefiniteness in Latin as well. Generally, what is new to a discourse cannot be assumed to be identifiable and so is indefinite. Indefiniteness might also be the best translation of an ‘undifferentiated element of a larger group or set.’ This is in concord with Lyons’ notion of indefiniteness as cardinality, the indication of a number or amount. Latin has a number of cardinal expressions, words for ideas like ‘any,’ ‘some,’ ‘all,’ ‘neither,’ 'few,' ‘most,’ etc. While Latin has no equivalent to the English ‘a/an,’ it does have a host of these cardinality terms, including the word ‘unum’ for ‘one.’ Eventually, this word would of course become the indefinite article in Romance (Sp. 'uno/una,’ Fr. ‘un/une,’ etc.).

Lyons brings out a host of information on how a grammarian might tackle the issue of definiteness. However, I believe he misses much of the point of definiteness. Since definiteness entails the speaker making a judgment of the hearer’s state of knowledge, and since the hearer must cognitively respond to this judgment and find the referent, definiteness can be understood as tying together the participants of a discourse. This phatic function, by which a channel is established through tying minds together in talk, unifies people under a communicative umbrella. Humans are unique among the animals in being able to refer to what is NOT part of the here and now. Yet shifters and indexicals show through their very definiteness that the here and now of discourse is an important arena for both semantic negotiation and social action. Many languages lack definiteness, but all seem to find ways of cognitively and socially tying people together, indicating that the participants of the talk are all aligned and on the same page. Since definiteness, whether defined syntactically, semantically, or pragmatically, is not universal, we must broaden the scope of the inquiry to see how languages, both written and spoken, tie participants together. Lyons’ big book of THE is certainly big enough already, but it could perhaps be better. What is needed is not more about determiners in English-like languages, but rather an understanding of how grammatical categories merge or split or simply bleed into one another over time and across space. Universal Grammar cannot be salvaged here. If the idea of innate lexical and functional categories must be jettisoned, we best look to the functions of grammar to see how formal variation can be brought into line with what we know about humans and their talk.

[Rebecka · Rating 4 out of 5]

Even though I haven't read the final part of this book, I'm finally going to mark it as read, since it is more of an encyclopedia than a book that necessarily needs to be read in its totality. I use this book frequently for my master's thesis, and even though parts of it are quite heavy, it's really a good read. My only problem with it is that it is *so* long, so extensive, and I sometimes get lost in it. For anyone working with definiteness, it's obviously a must-read.