An accident is that which is none of these, not a definition, property or genus, but it belongs to the thing, and which is admitted to belong to one and the same thing, and also not belong, such as “sitting” is admitted as belonging to one and the same thing and also not belonging. Likewise “whiteness” is also an example. There is nothing to hinder the same thing being at one time white and another time not-white. The second of the definitions of accident is better. For with the first definition being said, it is necessary, if someone is going to understand, to know beforehand what a definition and genus and property are, but the second is self-sufficient for knowing what the thing said is by itself. The things compared with one another are also placed as among the category of “accident,” if they are spoken of in any way owing to the accident, such as whether the noble or fitting is more choice-worthy, and whether the life of virtue or pleasure is more pleasing, and if anything else happens to be said resembling these questions. In all these cases an investigation begins as to which of the two does the predicate preferably belong to. And it is evident from these that nothing hinders the accident from becoming a property in relation to something else, such as “sitting” is an accident, whenever someone sits, but it will be at another time a property, namely the property of not being the only one sitting in relation to others not sitting. The result is that nothing hinders an accident from becoming a property in relation to a certain thing or time. But it will never be a pure property.1

Topica 102b4-b26

In this section Aristotle offers two definitions for accident. The first is that an accident is not a definition, property or genus. This is problematic, as he admits, for one must first know what these later three are before one can create a conceptual category that accommodates their negation. His clearer definition is that an accident is something, as an attribute or property, which can apply or not apply to something at different times. This definition tracks well with our English understanding of accident, for instance when we say that something is ‘accidental,’ we mean that things could have turned out differently than they did. There is no necessary correspondence between what could have happened and what did happen.

Also related to accidents are questions regarding the assignment of an accident, such as, “Does x or y have more of z?” In Aristotle’s examples, he asks which of two things are “more choice-worthy” or “more pleasant.” Of the “choice-worthy,” he asks to which of two types of life, the life of virtue or the life of leisure, is the term properly predicated. Given our external knowledge of Aristotle, such as that the life of virtue is indisputably more choice-worthy than the life of leisure (viz. Nicomachean Ethics) we can assume that Aristotle is making a grammatical point here, more so than a philosophical one. This is further corroborated by the fact that he follows this up by saying, “In all these cases an investigation begins as to which of the two does the predicate (τὸ κατηγορούμενον) preferably belong to.” I take Aristotle then, to mean that the assignment of an accident– as the predicate of a sentence– also occurs when there is a comparison of terms where one thing has more or less of the accident than another thing. For example, “Is the shoe or pencil heavier?,” has the answer, “The shoe.” Thus we say, “The shoe is heavier.” In so far as this can be said, it is said of the shoe accidentally, as this could change to, “The shoe is is not heavier,” if we were to ask, “Is the shoe or chair heavier?”

Returning to his previous definition of property, Aristotle notes that some accidents can be a property. But reiterating his previous distinction between real properties and quasi or contingent properties (see 102a18), he informs us that an accident can only be of the former type.

1 Συμβεβηκὸς δέ ἐστιν ὃ μηδὲν μὲν τούτων ἐστί, μήτε
ὅρος μήτε ἴδιον μήτε γένος, ὑπάρχει δὲ τῷ πράγματι, (5)
καὶ ὃ ἐνδέχεται ὑπάρχειν ὁτῳοῦν ἑνὶ καὶ τῷ αὐτῷ καὶ μὴ
ὑπάρχειν· οἷον τὸ καθῆσθαι ἐνδέχεται ὑπάρχειν τινὶ τῷ
αὐτῷ καὶ μὴ ὑπάρχειν· ὁμοίως δὲ καὶ τὸ λευκόν· τὸ γὰρ
αὐτὸ οὐθὲν κωλύει ὁτὲ μὲν λευκὸν ὁτὲ δὲ μὴ λευκὸν εἶναι.
ἔστι δὲ τῶν τοῦ συμβεβηκότος ὁρισμῶν ὁ δεύτερος βελτίων· (10)
τοῦ μὲν γὰρ πρώτου ῥηθέντος ἀναγκαῖον, εἰ μέλλει τις συν-
ήσειν, προειδέναι τί ἐστιν ὅρος καὶ ἴδιον καὶ γένος· ὁ δὲ
δεύτερος αὐτοτελής ἐστι πρὸς τὸ γνωρίζειν τί ποτ’ ἐστὶ τὸ
λεγόμενον καθ’ αὑτό. προσκείθωσαν δὲ τῷ συμβεβηκότι
καὶ αἱ πρὸς ἄλληλα συγκρίσεις ὁπωσοῦν ἀπὸ τοῦ συμ- (15)
βεβηκότος λεγόμεναι, οἷον πότερον τὸ καλὸν ἢ τὸ συμφέρον
αἱρετώτερον, καὶ πότερον ὁ κατ’ ἀρετὴν ἢ ὁ κατ’ ἀπόλαυ-
σιν ἡδίων βίος, καὶ εἴ τι ἄλλο παραπλησίως τυγχάνει τού-
τοις λεγόμενον· ἐπὶ πάντων γὰρ τῶν τοιούτων ποτέρῳ μᾶλ-
λον τὸ κατηγορούμενον συμβέβηκεν ἡ ζήτησις γίνεται. δῆ- (20)
λον δ’ ἐξ αὐτῶν ὅτι τὸ συμβεβηκὸς οὐθὲν κωλύει ποτὲ καὶ
πρός τι ἴδιον γίνεσθαι· οἷον τὸ καθῆσθαι, συμβεβηκὸς ὄν,
ὅταν τις μόνος καθῆται, τότε ἴδιον ἔσται, μὴ μόνου δὲ καθ-
ημένου πρὸς τοὺς μὴ καθημένους ἴδιον. ὥστε καὶ πρός τι καὶ
ποτὲ οὐθὲν κωλύει τὸ συμβεβηκὸς ἴδιον γίνεσθαι. ἁπλῶς δ’ (25)
ἴδιον οὐκ ἔσται.