Been attending lectures to keep myself busy until I leave next month, and I can't seem to get away from Aristotle, even in courses which theoretically have nothing to do with him. So I figure it's a good time to resume my series of reflections on language in Aristotle.
Physica B 3. 194b 16 – 195a 3
Διωρισμένων δὲ τούτων ἐπισκεπτέον περὶ τῶν αἰτίων, ποῖάτε καὶ πόσα τὸν ἀριθμόν ἐστιν. ἐπεὶ γὰρ τοῦ εἰδέναι χάριν ἡ πραγματεία, εἰδέναι δὲ οὐ πρότερον οἰόμεθα ἕκαστον πρὶν ἂν λάβωμεν τὸ διὰ τί περὶ ἕκαστον (τοῦτο δ’ ἐστὶ τὸ λαβεῖν τὴν πρώτην αἰτίαν), δῆλον ὅτι καὶ ἡμῖν τοῦτο ποιητέον καὶ περὶ γενέσεως καὶ φθορᾶς καὶ πάσης τῆς φυσικῆς μεταβολῆς, ὅπως εἰδότες αὐτῶν τὰς ἀρχὰς ἀνάγειν εἰς αὐτὰς πειρώμεθα τῶν ζητουμένων ἕκαστον. ἕνα μὲν οὖν τρόπον αἴτιον λέγεται τὸ ἐξ οὗ γίγνεταί τι ἐνυπάρχοντος, οἷον ὁ χαλκὸς τοῦ ἀνδριάντος καὶ ὁ ἄργυρος τῆς φιάλης καὶ τὰ τούτων γένη· ἄλλον δὲ τὸ εἶδος καὶ τὸ παράδειγμα, τοῦτο δ’ ἐστὶν ὁ λόγος ὁ τοῦ τί ἦν εἶναι καὶ τὰ τούτου γένη (οἷον τοῦ διὰ πασῶν τὰ δύο πρὸς ἕν, καὶ ὅλως ὁ ἀριθμός) καὶ τὰ μέρη τὰ ἐν τῷ λόγῳ. ἔτι ὅθεν ἡ ἀρχὴ τῆς μεταβολῆς ἡ πρώτη ἢ τῆς ἠρεμήσεως, οἷον ὁ βουλεύσας αἴτιος, καὶ ὁ πατηρ τοῦ τέκνου, καὶ ὅλως τὸ ποιοῦν τοῠ ποιουμένου καὶ τὸ μεταβάλλον τοῦ μεταβαλλομένου. ἔτι ὡς τὸ τέλος· τοῦτο δ’ ἐστὶν τὸ οὗ ἕνεκα, οἷον τοῦ περιπατεῖν ἡ ὑγίεια· διὰ τί γὰρ περιπατεῖ; φαμέν «ἵνα ὑγιαίνῃ», καὶ εἰπόντες οὕτως οἰόμεθα ἀποδεδωκέναι τὸ αἴτιον. καὶ ὅσα δὴ κινήσαντος ἄλλου μεταξὺ γίγνεται τοῠ τέλους, οἷον τῆς ὑγιείας ἡ ἰσχνασία ἢ ἡ κάθαρσις ἢ τὰ φάρμακα ἢ τὰ ὄργανα· πάντα γὰρ ταῦτα τοῦ τέλους ἕνεκά ἐστιν, διαφέρει δὲ ἀλλήλων ὡς ὄντα τὰ μὲν ἔργα τὰ δ’ ὄργανα.
Having defined these things, one must give consideration to the αἰτίαι, of what sort and how many there are. For since our investigation is for the sake of knowing, and we do not think we know each thing before we grasp the ‘why’ [διὰ τί] concerning it, which means to grasp the primary αἰτία, it is clear that one must do this for creation and decay and all natural change, so that knowing the principles of these things we may try to refer each thing which we seek to them.  In one way the thing out of which something comes to be is said to be responsible [αἴτιον] for it; thus bronze for a statue, and silver for a bowl, and things of that type.  In another way the form and the pattern is responsible [αἴτιον]; that is, the account of [a thing] being what it is and things of this type (thus [the αἰτία] of an octave [is] two in relation to one, or number in general) and the parts in the account.  Additionally, [an αἰτία is that] from which the first principle of change or stillness [comes]; thus one who advises is responsible [αἴτιος], and the father [is responsible for] the son, and generally, the making for the thing being made and the changer for the thing changed.  Further, [an αἰτία is] like an end [τέλος], that is, the thing for the sake of which; thus, health [is an αἰτία] of walking around. For what does he walk? We say, so that he may be healthy, and saying this we think that we have explained its cause [αἴτιον]. And such are all things which, once something else has begun the motion, meanwhile become [something] for the end.
αἴτιον, αἰτία: originally “responsibility, blame, accusation”, related to αἰτέω “to ask, beg”. In the Hellenistic period it also has the meaning of a story explaining the origins of a custom, which is fairly close to how Aristotle uses it here. It governs the genitive.
‘Αἰτία’ has no good English equivalent. The usual translation is “cause”, and the four αἰτίαι Aristotle lists are conventionally known as  material cause,  formal cause,  efficient cause, and  final cause. However, this translation is unsatisfactory for a number of reasons which will become clear. While the Latin causa is similar in meaning to the Greek, ‘cause’ in English is much more restricted. It implies an event, or at least an agent which makes something else happen; the only one of Aristotle’s causes which makes sense in this meaning is efficient cause.
“Reason” or “explanation” (if understood as referring to things, not words) would not be inappropriate, but are often required to translate the Greek λόγος. “Things because of which” is closer to Aristotle’s meaning, but awkward. I have translated the adjective as “responsible” where sense permitted; elsewhere I have used the Greek word.
Moravcsik, who points out that Aristotle’s αἰτίαι are typically not events or propositions, but things, argues that this passage should be interpreted as a theory about understanding (what is necessary in order to ‘know’ something), and that the translation “cause” is therefore highly misleading. Instead, Moravcsik argues, the αἰτίαι should be called “generative factors,” meaning that they account for something being the type of thing it is: “for x to be an aitia of y is for x to be in a relation to y such that the grasp of that relation enables one to understand some important aspect of y” (624). The theory suggests that, although the specific account of each individual thing differs, it is possible to consistently identify certain types of relationships – Moravcsik’s phrase is “configurations in reality” – which explain what a thing is.
Pustejovsky, influenced by Moravcsik’s ideas, incorporated Aristotle’s αἰτίαι into a theory of lexical semantics. His Generative Lexicon involves representing words according to several categories which describe both combination rules and meaning components. One of them, Qualia Structure, which describes the properties of a lexical item, is taken from Aristotle and broken down into categories which correspond with the four αἰτίαι: constitutive, formal, telic, and agentive (Saeed 281). The approach is heavily computational, concerned primarily with identifying sets of elements which would allow us to construct the meaning of any sentence.
It seems to me, however, that the αἰτίαι, as an analytic system which differs significantly from modern logic, potentially have other applications in semantics. When we use words to describe a thing, we pick out certain aspects of a thing which we wish to emphasize; when asked who someone is, for example, we may reply “Polyclitus” or “my neighbor” or “the sculptor” depending on the situation. What the αἰτίαι offer is a way of examining which elements are being picked out and what type of information they are being used to convey.
Additionally, αἰτίαι seem to play a role in how we are able to use words in extended senses. Although Moravcsik suggests that the relationships described by αἰτίαι are ontological, not linguistic, there seems to be some awareness of these relationships in the ways we use language. In particular, the αἰτίαι seem to be relevant to the problem of polysemy, as Pustejovsky recognizes but perhaps does not take far enough, as he does not seem to have investigated how different uses of words relate to their αἰτίαι structure. Are certain αἰτίαι present when we talk about Athena the goddess that are different when we talk about a statue of her?
Moravcsik, J. M. 1975: Aitia as Generative Factor in Aristotle's Philosophy. Dialogue 14, 622-36.
Pustejovsky, James. 1995: The Generative Lexicon. Cambridge, MA: MIT Press.