Ancient Greek Philosophy
Reference in the philosophy of language is, at a minimum, a relationship between words and items in the world, typically individual objects, but depending on oneās views about what there is, words can also refer to events, processes, classes, or kinds. This rudimentary notion of reference is found in Pre-Socratic philosophical fragments. Parmenides of Elea (c. 475 BCE) refers to the custom āof namingā and that human beings have names for both āBecoming and Perishing, Being and Not-Beingā, although he believes it is a mistake to have names for Not-Being because what is not the case is neither thinkable nor expressible, and thus these names are unintelligible (Freeman 1983, 43ā44). The 5th-century Sophist Gorgias of Leontini, at least as it was reported by Sextus Empiricus (c. 200 BCE), distinguished human speech from things which are external to human beings and concluded from this that language cannot give any information about the external world because speech can only be about mental states. So Parmenides, on the one hand, identified the intelligibility of language with reference to existing things, making it impossible to literally say something false. On the other hand, Gorgias cannot understand how a word can refer to something else, and thus for him it is impossible even to say something true about the world.
Plato (429ā347 BCE) avoids these absurd consequences by distinguishing the intelligibility of language from reference. How language represents what is independent of language and its users was a focal point for Plato (429ā347 BCE) in his dialogue Cratylus. Plato considered two positions: (a) names refer to objects on account of conventions that stipulate that the name refers to that object, and (b) names refer to objects when the name naturally has or displays the nature of the object, either as a likeness or as a description of the object. He rejected both accounts and argued that naming depends on having prior knowledge of the named objects, suggesting that the intelligibility of names rests on something distinct from either the name or object. Plato identifies the human understanding as the source of the intelligibility of names.
Plato maintains that human conceptions of how things are can be incorrect while still being intelligible, and consequently, human beings can speak falsely as well as truly. In the later dialogue, Sophist, Plato returns to this topic and argues that meaningful speech can be true or false, rejecting the Parmenidean view that meaningful but false speech is impossible. True speech, Plato maintains, says about things āas they areā and false speech āsays things different from those that areā (263aāb). This is possible because meaningful speech is a āweaving [of] verbs with namesā and these can be woven together to make different statements about how things are (262d). By distinguishing between how something is said to be and how things are, Plato begins to differentiate meaning and reference.
Platoās student Aristotle (384ā322 BCE) characterizes the intelligibility of human speech in terms of three components: spoken sounds, affections of the soul, and actual things. Spoken sounds signify affections of the soul, which include both thoughts and perceptions. Affections of the soul, in turn, enable human beings to think about actual objects in virtue of the human understanding having forms in common with these objects. The relationship between speech and affections of the soul is conventional, but the relationship between thought and actual things is natural because whether or not a thought or perception has forms in common with actual things is not something that can be stipulated. Language is intelligible because it signifies affections of the soul, and affections of the soul are intrinsically intelligible because forms for Aristotle, as was the case for Plato, just are the basic sources of intelligibility.
The meaning of speech, then, is a function of the signified affections, but this is not a referential relationship. Reference, as a relation between words and things, holds between spoken words and actual things when the forms of the signified thought or perception are also forms of actual things. While spoken words can refer to actual things via the signified affections of the soul, the affections themselves do not refer to actual things. Thoughts or perceptions of actual things display the forms of actual things and thus actual things and affections literally have something in common, namely the same forms, but this does not mean that thoughts and perceptions of actual things are identical with actual things. For example, the perception of a stone is not identical to the stone nor is the stone itself present in the perception, but the form of the stone, without its matter, is in the perception (De Anima, 431b20ā30).
Aristotle does not have a special term for reference, but he explicitly distinguishes signification from the relation to actual objects. There is no such thing as a goat-stag, but nevertheless, Aristotle writes, āgoat-stagā signifies something (De Interpretatione, 16a16ā17). Languageās relation to actual things becomes relevant for Aristotle when considering the truth or falsity of sentences that make statements, namely sentences that are used to affirm or deny something of something. While āgoat-stagā by itself signifies a form, by itself it is neither true nor false because it does not affirm or deny anything. However, once we add a verb, for instance, āisā, āwill beā, or āwasā, and make a statement, whether or not there is an actual goat-stag is relevant. The affirmation āA goat-stag isā, where āisā is not a copula but predicates existence, is false because there are no goat-stags. On the other hand, the affirmation āA man isā is true because an actual thing has the form of man. The difference is that the form of goat-stag has no matter and the thought and object are identical, but the form of man that is displayed in thought without matter is in reality combined with matter to form an actual thing.
In sum, for Aristotle, to evaluate the intelligibility of speech, affections of the soul need to be considered, but when evaluating the truth or falsity of statements, the relationship of speech to actual things needs to be considered. Nevertheless, although the reference relation between words and things is recognized by Aristotle and assigned a role in his philosophy of language, this is not a central topic for him. His primary concern regarding language was meaning and the logical relationships of statements, which he understood to be independent of the question of what there is. That was the topic of metaphysics.
The concern with meaning abstracted from reference also prevails in Stoic philosophy of language. The Stoics distinguished between the sound-forms of speech (phÅnÄ), written forms (lexis), what an utterance signifies or expresses (lekton), and the unity of sound and meaning, meaningful speech (logos). Stoic philosophy of language sharply distinguished lekta, or what is expressed with language, and the extra-mental objects or events to which meaningful speech refers. An important property of lekta is that they are expressible in language, but they are immaterial abstractions that are not identified with occurrent psychological states or with the sentences that express them. Truth-value is a property of lekta that form statements, not the sentences that express them (Diogenes 1925, 57ā66; Bobzien 2003, 85ā88; Law 2003, 38ā42). The concept of something expressible in language, but distinct from language, mind, and what exists plays a central role in medieval philosophy and prefigures the notion of propositional content in the philosophy of language of Gottlob Frege (1848ā1925).
Medieval Philosophy
Early medieval philosophy of language continues the trajectory set by Aristotle and the Stoics to focus on propositional content and its logical structure, abstracted from reference, as the ground of the intelligibility of language. Building on the Stoic concept of the lekton, Augustine of Hippo (354ā430) in his De Dialectica distinguished between an utterance (verbum) and the utterance used to say or express something (dictio). What is said or expressed is something sayable or expressible (dicibile), which is not something heard or seen, but is understood by and contained in the mind (Augustine 1975, v:88). What is expressible has a status independent of the actual expression because it can be understood by a mind independently of the expression. Augustine identified a fourth component of speech, namely the object that is signified by speech (res), but the intelligibility of language depends on dicta, not res.
While Augustine isolated the component of what is āsayable [dicibile]ā, Peter Abelard (1079ā1142) analyzes what is sayable into distinct components. Abelard distinguished the propositional content of a statement (enuntiatio, dictum propositionis) from the affirmation expressed by the statement. The content of a statement is what is understood or proposed when spoken, and Abelard classifies this as a kind of signification (significatio intellectualis), not to be confused with the signification of objects (significatio realis) (Abelard 1956, 148, 154ā156). What is understood, the propositional content, can be shared by other kinds of speech acts beside statements, such as commands, questions or wishes (Abelard 1956, 151). A statement is an assertion or affirmation of what is understood, and as such it is either true or false (Abelard 1956, 154). So truth-value is not a property of content, but the affirmation of content. By separating propositional content from the assertion of the content, Abelard was able develop a better understanding of sentential connectives and operators, motivating the study of what came to be known as the syncategorematic terms.
Abelard calls reference understood as a word-world relationship ānominatioā or āappellatioā to distinguish that relationship from the relation of significatio. Nominatio or appellatio do not play a role in determining the propositions that language expresses. In other words, reference does not play a role in determining the semantics of natural language, and hence Abelard believes that reference can be neglected in the study of the meaning of language (Abelard 1984, 233/308.19ā22). The primary semantic function of names is not to refer to something, but to signify concepts. The role for reference for language is to determine whether an affirmation of a proposition is true or false. While this is essential for knowing how the world is, for Abelard this is not essential for understanding natural language.
Reference begins to be recognized as having a role in determining the meaning of natural language with the introduction of supposition theory in 13th- and 14th-century European philosophy of language. Supposition theory had two taproots: (i) the distinction between signification (significatio) and naming or calling (nominatio or appellatio), already developed by Abelard, and (ii) the concept of a suppositum, namely of something that underlies or instantiates the signified concept or nature, which already appears in Boethiusās commentaries on Aristotle. Suppositum also had a grammatical or syntactic use, namely to be the subject of predication in a statement. Combined, these concepts suggest a new area of inquiry: what objects are named by the noun that is the grammatical subject of a statement?
Medieval grammarians, logicians, and philosophers turned to this issue in response to the intensified study of fallacies and sophisms that the rediscovery in Europe in the 12th century of Aristotleās Sophistical Refutations inspired. A widely shared diagnosis of fallacies and sophisms was that they resulted from not recognizing that while the signification of a noun can remain the same, in different contexts of statements, the objects that are named by the noun can change depending on what is being predicated in a statement. For example, Peter of Spain (d. 1277) in his widely used introduction to logic textbook Summulae Logicales maintains that while the signification of āhuman beingā is constant, there is a difference between the statements āA human being is runningā and āHuman being is a speciesā. The failure to recognize a difference leads to fallacies. While inferring āSome human being is runningā from āA human being is runningā is valid, inferring āSome human being is a speciesā from āHuman being is a speciesā is a fallacy (Peter of Spain 1972, 69ā72).
Depending on the philosopher, what is signified by a term is a common nature, essential property, substantial form, defining concept, or a class of past, present, future, and possible individuals. While what a word signifies can change by imposing a new signification on a wordāfor example, when a word is redefinedāit was assumed that given that a word has a signification, there are no changes in what it signifies. This assumption was grounded in the view that a term signifies a common nature or substantial form and these were assumed to be changeless. So the difference had to be located elsewhere, a...
