The contents of this chapter have been removed for rights reasons.

Philosophers have come to speak of an Augustinian picture of language. The picture is not really Augustine's, as we shall see, but it makes a good starting point for exploring what his views actually were. Those views, though not as crude as the “Augustinian picture,” will turn out to be mainly unoriginal, following a tradition that was already several hundred years old in his day, and helping to sustain that tradition for a further millennium and more. We know better now, thanks mainly to the fundamental insights of Frege and Wittgenstein.

In writing De Trinitate Augustine had three main objectives. He wished to demonstrate to critics of the Nicene creed that the divinity and co-equality of Father, Son, and Holy Spirit are rooted in scripture. He intended to tell pagan philosophers the need for faith in a divine mediator so that divine self-revelation and redemption can occur. Finally, he wanted to convince his readers that salvation and spiritual growth are connected with knowing themselves as images of the Triune God, from whom they came and toward whom they go, with a dynamic tendency to union realized by likeness to God who is Love.

That a spoken sound is significant is a function of its having an internal correlate in the mind. In semantic contexts, Aristotle identifies the mental state with a logos treated as a meaning or a specification of meaning in a definition. This condition on meaningful sounds is augmented by the requirement that true sentences must also correspond to extralinguistic states of affairs. A true sentence is reducible to an internalized logos, its meaning, since the sounds and/or graphic representations of it are signs of the mental state. Thus the correlation between the internal content and the external object specified by the theory of meaning becomes a threat to the existence of knowledge – or at least to our having any confidence that any particular sentence or body of assertions, such as those making up a demonstrative science, are true. Aristotle seeks to defuse this threat by in effect contextualizing the two relevant aspects of meaning, the pathema as a logos that is expressible in a natural language and the pathema as a likeness to an extralinguistic object. The context in which these are considered in the Posterior Analytics is the theory of demonstrative science, specifically, the theory of definition and the account of the origin of universals. The basic sentences or premises of a demonstrative science are true ex hypothesi. A demonstrative science is a deductive science, moreover, and this means that definitions of basic objects will be included among the primary premises.

In De Interpretatione 1, no sooner has Aristotle asserted that the mental states that words stand for are likenesses of the pragmata of the world, than he dismisses this topic with the words:

The place in the De Anima that Aristotle has in mind in the De Interpretatione is not easily discovered. Nor is it entirely clear which features of the theory of meaning will or should be explained there. Is it merely the nature of the psychic pathemata (internal states) or the likeness obtaining between the pathemata and their external correlates (pragmata)?

Speech is not mentioned in its own right in the De Anima but only in connection with hearing.

Here the intent is to differentiate between speech and other types of noises produced by the respiratory system. What the passage adds to the De Interpretatione's description is the explicit mention of imagination (τασία).

No sooner has Aristotle summarized his theory of meaning in the De Interpretatione than he turns to the topic of sentences. Sentences are the basic linguistic item to which truth values can be assigned, and they are made up of linguistic units that have meaning. To avoid the problems of the Cratylus, Aristotle draws a sharp distinction between the basic units of meaning and the basic truth-functional entities. He also acknowledges that language has functions other than assertion. That said, sentences have truth values in virtue of what is expressed by the sentence, and this is a function of the meaning of the parts of a sentence. A sentential part has a meaning just in case it is a conventional sign for a thought, the content of which is determined by its relation to an object that is external to the thought. Combining terms to make an assertion yields a statement that expresses a nexus of external features. Whether the assertion is true or not depends upon whether this nexus obtains in fact. Certain peculiarities of Aristotle's discussion of truth – more accurately described as a discussion of falsity – fall into place when one recognizes that Aristotle's starting point is a conception of truth-functional entities that are compounds. These compounds are constructed from meaningful (but not independently truth-functional) linguistic units. Similarly, certain peculiarities of his conception of necessity fall into place when it is recognized that for Aristotle the force of the necessity predicate is to indicate that the nexus of features always obtains.

This book began as a paper on De Interpretatione 1, 16a3–8:

This description is Aristotle's only explicit attempt to define meaning, and it has been, as a recent commentator remarked, “the most influential text in the history of semantics.” Notwithstanding, the account of meaning found in these lines was dismissed by John Ackrill as inadequate, and Aristotle has found few defenders in more recent literature. Suspecting that the negative assessment had been too hasty, I set out to discover whether Aristotle might, after all, be right about meaning. Framing the issue in this way proved jejune, and my initial query was replaced by a series of questions: What position is Aristotle taking here? Is it an account of meaning in the modern sense? Is this account one that he tries out in a relatively early work and later rejects? Does the conception of language embodied in these lines help us better understand the linguistic and ontological notions of definition and universals that are key players in Aristotle's epistemology and metaphysics?

Granted, the theory of meaning expressed in these lines is puzzling. Two troublesome relations are posited – one between a significant sound and an internal state and the other between that state and the external object of the state.

The ontological requirements of language and knowledge, as Aristotle understands them, are such that enduring characteristics of mutable and transient individuals are needed. This prompts Aristotle to identify forms, in contrast to concrete particulars, with primary substance in the central books of the Metaphysics and to develop an account of the relation between form and matter that is informed at every stage by a concern with linguistic definition and a conception of essence as the exemplary object of knowledge. A crucial concept in this effort to knit together several objectives is λόγоζ (logos). The term logos is used by Aristotle in many contexts and for many things, from strings of symbols to forms, from speech to definition. In Metaphysics VII, Aristotle exploits the ambiguity of logos to display the common feature of a thought and its referent in the world. The logos actualized in matter is a composite substance, and the same logos realized in thought is a meaning. This is a puzzling doctrine but also by this point in our inquiry a familiar one. Not only does Aristotle assign logos both an ontological role and a linguistic one, he also uses horismos (definition in the strict sense) and pathema (vehicle of meaning) to indicate linguistic items and ontological ones. The real definition is a feature of the world, not just a specification of a meaning; the pathema is both a state of mind and its content (a meaning).

At every stage, the description of ontology is guided by the requirement of rational accessibility. Not only must the ontology be intelligible from a theoretical perspective, it must be accessible at the pre-theoretical level of perception and ordinary language.

The most basic existents in the Categories are particular substances and particular attributes. Meaning in the De Interpretatione is a function of reference to existents. Definitions of the basic objects of a demonstrative science must express real natures, according to the Posterior Analytics. Equally as much as science, language depends upon the apprehension of universals. Linguistic and scientific definitions articulate universal concepts. There is an apparent tension between the role of universals in Aristotle's epistemology and semantics and the Categories doctrine that universals are secondary and dependent upon human classificatory schemes. This is one problem Aristotle faces in the Metaphysics.

Another problem facing Aristotle is how to explain the relation between the meanings of the terms of natural language and the meanings of terms in scientific discourse (often the same linguistic sign has both uses.) Science requires meanings that correspond precisely to reality. Natural language requires words (at least designative ones) that refer to real existents. There can be no doubt that Aristotle distinguishes between the two cases. The nominal definition of a natural language term may express a concept that has no exemplifications, for instance, ‘goatstag’, or it may embody an imperfect understanding of the real existent, for instance, ‘human.’ What, then, is the difference between a meaning expressed in a linguistic definition and a meaning expressed in a definition of essence? This concern motivates the discussion of definition and essence in Metaphysics VII.

How a mind grasps universals is addressed all too quickly in Posterior Analytics II 19. The challenge facing Aristotle is to fill out this quick account in a way that makes the acquisition of universals intelligible. This would be a challenge under any conditions; however, the importance of meeting this challenge is enormous for Aristotle, whose logic and theory of demonstrative science recognize only inferences through universals. The theory of meaning in the De Interpretatione does not identify significant sounds with universal terms; however, universal terms are much more common in ordinary discourse than are proper names for individuals, and Aristotle's claim that the mental state is the same for all humans who share a concept supports the identification of meanings with universals. Finally, and perhaps most significantly, knowledge requires universals. On this point, Aristotle (as we have seen) is in complete agreement with Plato. He differs in making experience the basis for the knowledge of universals. Posterior Analytics II 19 sketches the process involved, making no distinction between the universals required for ordinary discourse and those required for science, or between the universals required for one type of science or another. Elsewhere, Aristotle recognizes three types of theoretical science (mathematics, physical science, and metaphysics). These differ as their objects differ with respect to primacy, universality, and intelligibility. Posterior Analytics II 19 would lead us to expect that universals having the epistemic characteristics required by a particular science would be grasped through the inductive process (epagoge). Conversely, if this is not the case, the claim to generality of the Posterior Analytics account of first principles would be undermined.