Language, Philosophy of: Ancient and Medieval

Only in recent times has philosophy of language been considered a distinct branch of philosophy. But ancient and medieval philosophers had different, sophisticated theories about the relation between language—both individual words and whole sentences—and reality, and the thirteenth century saw one of the most thorough attempts ever to give an abstract, analytical account of the grammar of a natural language.

Words and Things: Plato and Aristotle

In the Cratylus, the one dialogue he devoted exclusively to questions about language, Plato (c. 428–348 or 347 b.c.e.) contrasts two ways of explaining how words link with things. Is it purely a matter of "convention and agreement," so that "whatever name you give to something is the right one"? Or are some words naturally suited to stand for certain given things? Superficially, Plato certainly seems to lend support to the latter, naturalist answer, exploring through usually fanciful etymology how words can be analyzed into significant elements (for instance, psuchos —soul—"derives" from echei —has/holds—and phusin —nature), and even attributing aspects of meaning to the sounds of individual letters. Yet Plato also points, by irony and more directly, to the inadequacies of such naturalism. Other philosophical schools, such as the Stoics and Epicureans, held to a naturalist view without such reservations.

But not Aristotle (384–322 b.c.e.). He was clearly a conventionalist: a name is "a spoken sound significant by convention." But how do such spoken sounds link with the world? In On Interpretation (16a), Aristotle writes that

Spoken sounds are symbols of affections in the soul, and written marks symbols of spoken sounds. But what these are in the first place signs of—affections of the soul—are the same for all; and what these affections are likenesses of—actual things—are also the same.

Aristotle's semantic scheme, in which things are signified by words only through the intermediary of thoughts or mental images ("affections of the soul"), is underwritten by his psychology and metaphysics. Human beings know about the world through being affected by the forms that account for things being as they are: the heat of a hot stone, for instance, or the humanity that makes someone a human. Aristotle is therefore justified, to his own way of thinking, in supposing that although words vary from language to language, the mental signs they stand for are the same for all people and can correspond directly with the objects themselves.

Sentences and Facts: Aristotle and the Stoics

Aristotle's aim in thinking about language was not merely to look at the relation of naming-words to things, but to explain how words combine to form assertoric sentences, which can be true or false. On Interpretation studies the functions of nouns and verbs and the mechanism of predication in some detail, but Aristotle remains rather vague about how sentences as a whole link up with reality. In the Categories (14b) he talks of what makes a sentence true (what we might call a state-of-affairs) as a pragma —a vague word, meaning act, thing, or matter.

In the following centuries, it was the Stoics who gave deeper consideration to the semantics of sentences. They distinguished between signifiers (words and groups of words as utterances: for instance Dion ), name-bearers (for instance, Dion himself), and significations. They called these significations lekta ("sayables") and understood by them the states-of-affairs revealed by utterances (for instance, Dion's standing, the lekton of "Dion stands"). From this description, it sounds as if the Stoics had an ontology that included both things and states-of-affairs, and the latter were called lekta because they neatly filled the role of being what assertoric sentences signify. The concept of lekta may well have originated in this way, but there are two important qualifications to consider. First, the Stoics held that not only complete assertoric sentences, but also other types of sentences, such as commands and questions, have their lekta ; and also that, in addition to these complete lekta, there are incomplete lekta, which are the meanings of predicates (for instance, standing, the lekton of "stands"). Second, the Stoics were materialists, and they considered lekta to be not merely incorporeal but not even to exist.

Abelard and the Early Middle Ages

Early medieval philosophers were deeply influenced by the semantic scheme of On Interpretation. The way in which Boethius (c. 480–c. 524) discussed it in his commentaries (early sixth century) linked with ideas they found in the writings of St. Augustine of Hippo (354–430) to suggest the following, widely accepted basic scheme: written language signifies spoken languages, which in turn signify a mental language common to all humans, and the terms of this mental language signify things in the world. In medieval texts signification is usually therefore a causal, psychological relation: w signifies a thing x if and only if w causes a thought of x in the mind of a competent speaker of the language. There were, however, other influences, especially the grammatical writings of Priscian (fl. 500 c.e.), who was heavily influenced by Stoic theories.

By the eleventh century, there was already a strong philosophical interest in questions about language. For example, St. Anselm of Canterbury (1033 or 1034–1109) wrote a dialogue about the problems caused by a word such as grammaticus, which means "grammarian" but also has the adjectival sense of "grammatical": is grammaticus a substance, then, or a quality? In Peter Abelard's (1079–?1144) logical writings (c. 1115–1125) the semantics of both words and sentences receive careful and searching attention. Abelard accepts the usual psychologico-causal understanding of signification, but his nominalism—the view that nothing except particulars exists—made it problematic. He accepts that predicates signify universals, and then (in accord with his theory) goes on to identify these universals with what are not things—mental images or, in the latest version of his theory, a thought-content. Abelard also developed an account of the semantics of sentences. Assertoric sentences signify what he calls dicta (literally, "things said")—by which he means states of affairs or (on some occasions, something nearer to) propositions. Dicta, however, are not things; they are, literally, nothing. The parallel with the Stoics' lekta is striking, although direct influence does not seem possible. Thirteenth-century thinkers talked of sentences signifying enuntiabilia (things able to be said), and some fourteenth-century thinkers use the term complexe significabilia (complexly signifiables)—in both cases bringing themselves even closer to the meaning of the Stoics' term.

Speculative grammar.

Speculative grammar was a striking, though short-lived, episode in medieval thinking about grammar. Its outstanding exponents, Boethius and Martin of Dacia, and Radulphus Brito, were arts masters at the University of Paris in the period from approximately 1250 to 1300. They aspired to give grammar the universality demanded of an Aristotelian science and, although they worked entirely with Latin, they believed that the underlying structure they were uncovering was that of any language, although each language represented it using different combinations of sounds.

At the basis of speculative grammar is the Aristotelian semantics, which aligns things in the world, thoughts, and words. The speculative grammarians held that there are modes of being (properties of things, such as being singular or plural, active or passive) and, parallel to these, modes of thinking (as when the intellect thinks of a thing as being singular or plural, and so on). The modes of signifiying (modi significandi— from whence the term modistic grammarians or modists ) parallel these modes of being and of thinking. So, a first imposition links a sound with a certain sort of thing, and this root becomes a part of speech by being given modi significandi that first of all make it into one of the parts of speech (noun, verb, and so on) and then add features such as case and number (for nouns) or tense and person (for verbs). The modists' assumption is that the Latin grammatical categories are precisely molded to the general structure of reality, which is captured accurately in thought.

The terminists and Ockham.

A little earlier in the thirteenth century, logicians at Paris and Oxford were busy developing a different approach to the relation between words and things, the theory of the properties of terms, which was given its most popular exposition in the Tractatus (the so-called Summulae logicales ) of Peter of Spain (Pope John XXI; d. 1277). The theory of the properties of terms concerns the way in which nouns refer in the context of a sentence. One distinction (in the terminology of the English logician William of Sherwood) is between material supposition, where a word refers to itself ("Man has three letters")—a medieval equivalent of quotation marks—and formal supposition, where it refers to something in the world. Formal supposition can be either simple or personal: in simple supposition, a word refers to a universal ("Man is a species "), whereas in personal supposition it refers to particulars in various different ways, which the terminists further distinguished. The theory provided for the (personal) supposition of a word to be "ampliated" or "restricted" by its context. By adding white to man, one restricts the supposition of man to just those men who are white; by making it the subject of a future-tense verb, one will restrict the supposition of the noun to men in the future. Expressions like "think of" and "it is possible that" ampliate the reference of nouns in their scope to include all thinkable or possible such individuals.

Terminism did not fit the interests of mid-and late-thirteenth-century thinkers, but it was revived again in the fourteenth century. William of Ockham (c. 1285–?1349) uses it, in an adapted form, in setting out one of the most elaborate medieval theories about mental language. William is so fully committed to the idea of our thoughts being naturally structured in a language-like (indeed, Latin-like) way, that he breaks the Aristotelian mold and holds that, rather than words signifying thoughts, both words (natural language) and thoughts (mental language) signify reality in much the same way. Since Ockham—like Abelard before him—was a nominalist, he could not accept the usual idea that simple supposition is of universals: according to him, a word has simple supposition when it supposits for a mental term, but does not signify it.

Linguistic Diversity: Dante and the Arabs

The mainstream Western medieval tradition of thought about language concentrated on a single language, Latin. By doing so, it gained the advantage of allowing philosophers to see more clearly the large, abstract questions that concern the relation of any language to the mind and to reality. It also suffered, because scholars (as appears strikingly in the case of the speculative grammarians) simply assumed that features in fact special to the structure of Latin were universal to every language. By the later Middle Ages, however, there was at least some analytical investigation of the facts of linguistic diversity. The great poet Dante Alighieri (1265–1321), writing his On Eloquence in the Vernacular in Latin to defend writing in Italian, tried to explain how all languages derive from the original tongue spoken by Adam, and how they have changed and developed.

Things were very different in the Arabic tradition. During the eighth and ninth centuries, a great quantity of Greek scientific and philosophical work was translated into Arabic. In the earliest period, at least, those thinkers in Islam who thought of themselves as followers of the Greek philosophers tended to play down the value of grammatical study. Their attitude led to a famous confrontation in 932 between the grammarian Abu Sa'id al-Sirafi and the philosopher Abu Bishr Matta. While Matta held that logic provided a universal key to thinking, al-Sirafi's contention was that Greek logic is based on the Greek language: writers in Arabic need, rather, to study their own language. The contrast with thirteenth-century Latin thought is piquant. There, grammar was made into a sort of universal linguistic logic; here logic itself is argued to be as particular as the grammars of different languages.

See also Language and Linguistics ; Language, Linguistics, and Literacy ; Philosophy: Historical Overview and Recent Developments .

bibliography

PRIMARY SOURCES

Aristotle. 'Categories' and 'De interpretatione.' Translated by J. L. Ackrill. Oxford: Oxford University Press, 1974.

Long, A. A., and D. N. Sedley. The Hellenistic Philosophers. 2 vols. Cambridge, U.K.: Cambridge University Press, 1987.

Peter of Spain. Language in Dispute: An English Translation of Peter of Spain's Tractatus.… Translated by Francis P. Dinneen. Amsterdam and Philadelphia, Benjamins, 1990.

SECONDARY SOURCES

Abed, S. B. "Language." In History of Islamic Philosophy, edited by Seyyed Hossein Nasr and Oliver Leaman, vol. 2, 898–925. London and New York: Routledge, 1996.

Barnes, J., and D. Schenkeveld. "Linguistics: 'Part II, 6.1'" In The Cambridge History of Hellenistic Philosophy, edited by Keimpe Algra et al., 177–216. Cambridge, U.K.: Cambridge University Press, 1999.

Ebbesen, S. "Language, Medieval Theories of." In Routledge Encylopedia of Philosophy, 10 vols, edited by Edward Craig, 389–404. New York: Routledge, 1998.

Everson, Stephen, ed. Language. Cambridge, U.K.: Cambridge University Press, 1994.

Rosier, Irène. La grammaire spéculative des modistes. Lille, France: Presses universitaires de Lille, 1983.

John Marenbon