“Deponent” is a Spurious Category

I’ve seen a few posts lately regarding the “problem” of deponency and/or the middle voice in ancient Greek. One blogger even suggests that we use a different word than “middle”, which is a dumb idea, because “middle voice” is a term of art, with a specific meaning that has only a tenuous relationship to the ordinary use of the word.

To a linguist, this is all very bemusing. Trying to build elaborate models and explanations to help English speakers wrap their minds around the idea that ancient Greek speakers used middle or passive constructions in contexts where English would use the active is just pandering to Anglo-centrism — all the models are attempting to explain Greek in terms of the writers’ English-language categories.

Look, folks, news-flash: ancient Greek is NOT English! The categories of ancient Greek are not those of English, and the ancient Greeks’ reasons for using a particular voice in a particular situation may simply be quite different from those of modern-day English-speakers.

And they may indeed have not had reasons! Far more of language is made up of arbitrary convention than most scholars of language would like to admit. A search for “reasons” (or “deep structure”, cough cough) is often at best an exercise in historical linguistics.

It might have been better had Greek been further grammatically from English — it’s hard to shoehorn an ergative-absolutive system, for example, into English-speakers’ conceptual framework — they just have to learn it on its own terms.

So in teaching ancient Greek it’s not a cop-out to say “that’s just how they did it”. The idea of “deponency” is actually a barrier to thinking in ancient Greek, because it tries to keep the learner using English concepts, instead of forming Greek concepts! I think sometimes language pedagogy goes overboard in trying to teach systems of rules. Languages are in general messy, and the most useful and interesting parts of language are often exceptions to the rules.

I’m brushing up on my Attic Greek right now by going through Reading Greek, which I cannot recommend highly enough, but for my own amusement, I’m not bothering with making sure I’ve got all the paradigms, or even memorizing new vocabulary. Of course, I did have the advantage of memorizing lots of paradigms back in school days, but I’m surprised at how much structure and vocab I’ve been picking up simply inductively. It helps that the texts are interesting, colourful and thus memorable.

10 thoughts on ““Deponent” is a Spurious Category”

  1. To a linguist, this is all very bemusing.

    What are we? Chopped suey?

    More specifically, this isn’t a question of Greek grammar vs. English Grammar, but a question of how to best describe Greek. The fact is that what is traditionally called voice in Greek really isn’t what most linguists call voice.

    To talk about us and this discussion with respect to Anglo-centrism is complete nonsense.

    1. Mike:

      What are we? Chopped suey?

      I think you misunderstand the referent. Let me rephrase: as someone published in the field of computational linguistics, I am bemused by the wrangling over voice in Greek.

      As the title of the post indicated, it’s about the invention of a whole new category (“deponent”) that exists neither in Greek nor in English, but is invented entirely to try to explain Greek grammar in the categories of English (“passive in form but active in meaning”, as if the Greeks couldn’t figure out their own grammar).

      The fact is that what is traditionally called voice in Greek really isn’t what most linguists call voice.

      Nonsense. Voice in linguistics is about the relationships between semantic actors (agent/patient, frex), and syntactic complements (subject/object, frex). Voice in Greek is a fine example of this (the subject of “erxomai,” for example, is obviously the patient, having been subjected to a change of state).

      The fact that scholars of Greek feel the need to “explain” those relationships in terms of those that hold in English (inventing the spurious category of “deponent” to do so) is what bemuses me.

  2. I’m glad you noticed my post on “Renaming the Greek Middle Voice.” If you had time to follow the discussion, you would see that I’m not actually advocating renaming it, but doing a better job of describing it. The post compares the Greek middle not only to English, but to Spanish and cites studies on other languages as well. What we’re doing over at grklinguist.wordpress.com is to clearly distinguish Greek from these other language, not make it seem more like them. In this respect you seem to have missed the point.

    If you are interested in the application of linguistics to Ancient Greek by a variety of linguists and well established Greek scholars, come back to visit us. You are welcome to join the discussion.

  3. Micheal: I’m afraid I only have the time to kibbitz once in a while, but thanks for the offer.

    Why make reference to any other language (“middle voice is kinda like Spanish reflexives”) at all? Why not develop a semantic model based purely on Greek evidence? Without using categories like “deponent”.

    For example, in the Greek middle voice, the subject is ambiguous between agent, patient or both. Which it is for a particular verb depends on convention (or “is lexicalized”), and may in fact not correspond to English, Spanish or Proto-Indo-European at all.

  4. Why make reference to any other language (“middle voice is kinda like Spanish reflexives”) at all? Why not develop a semantic model based purely on Greek evidence? Without using categories like “deponent”.

    Do you ever read language typology works?

  5. @Gordon
    You seem to assume that I approve of the category “deponent.” I’m not sure why. All of the discussion of “deponent” that has gone on over at Greek Language and Linguistics (grklinguist.wordpress.com) has been aimed at showing that “deponent” is both unnecessary and misleading. Your view of the middle voice, expressed in your example above, is essentially the same as what Carl Conrad has argued for some time. I agree totally.

  6. Gordon —

    The, then, is that you’ve connected a particular topic, deponency, to two posts that aren’t discussing deponency. Fundamentally, my post was dealing with the question of whether its fair to refer make a distinction between middle voice & passive voice, criticizing beginning grammars who make such a distinction, by pulling together various quotes from A. T. Robertson, a person who probably knew Koine Greek better than just about anyone else in at least 100 years.

    By linking to our posts, you have also connected what we wrote to a supposed search for reasons for why X is X in Greek. That is, again, something neither of us did. If you’re going to write a post about how deponency is an unhelpful category, more power to you! I’ll join in and write one myself, but to link to my post and then write about a different, but related, topic in a polemic matter is only going to confuse.

    The main question is: Who is this rant directed against? If it’s against Dr. Palmer’s post and my own, then it is fundamentally misplaced. Neither of our posts challenge such a claim about deponency and neither of us would be interested in challenge such a claim. We both would agree with you (though, I wouuld suggest that current semantic model you’ve put forward needs a good amount of nuance — a very niceLangenacker-esque sematnic network has been proposed by Suzanne Kemmer [Middle Voice, Linda Manny [Middle Voice in Modern Greek, & Rutgar Allan [Middle Voice in Ancient Greek]) — all of whom are best categorized as linguists.

    But if this particular isn’t directed toward our posts, then why in the world link to them???

    As for the question of voice and it’s validity as a grammatical category for Greek, you either need to read my comment slower or read more linguistic works.

    What is traditionally called voice in Greek isn’t what *most* linguists call voice:

    Standard Theory, Extended Standard Theory, P&P, G&B, and MP all treat grammatical voice as a syntactic process.

    Role & Reference Grammar treats grammatical voice as a syntax process.

    Relational Grammar treats voice as a syntactic process.

    Greek Voice cannot be treated as a syntactic process. You appear to know this, based on your statements above. So my statement holds true. Chomskian lingiustics in some form dominates linguistics across the board. Combine those with adherents of RRG & Relational grammar and I’d be willing to be that we have a majority.

    Frameworks that do voice in such a way that fits closer with Greek include: Lexical-Functional Grammar, Head-driven Phrase Structure Grammar, Cognitive Grammar, & Construction Grammar.

    PS – You’ve corresponded with my wife, who is currently writer her thesis on language & metaphor.

  7. The point of my post is in the title. The post was triggered by the phrase “the problematic issue of deponency” in the first sentence of Micheal’s post, which was linked to from Mike’s post.

    I submit that if one doesn’t wish one’s post to be about a thing, one shouldn’t mention it in the lede.

    Micheal: The things that I’m attacking are first: the fact that there is even still a “problem of deponency” (some people must still hold to it, if you feel the need to mention it, after all). Mastronarde and Reading Greek, both of which I’m using to learn Attic at the moment, still in this day and age contain awkward discussions of deponency, which fact is the second trigger for my post. In this we obviously agree.

    Second, however, is the idea of changing the terminology used to refer to the middle voice, to make it easier for English speakers to understand. This seems a recapitulation of the mistake made by those who introduced deponency — why try to explain Greek using English categories?

    Mike: you’re right: I should have said “any linguistic theory that is actually useful for describing anything different from Noam Chomsky’s idiolect of English.” Also, you forgot to mention Combinatory Categorial Grammar.

  8. I submit that if one doesn’t wish one’s post to be about a thing, one shouldn’t mention it in the lede.

    Granted, but your post was still misleading, because it (again, whether intentionally or not) characterizes both of us as finding the concept of deponency acceptable.

    As for the use of English categories. Middle isn’t an English category, so I’m not sure how it is relevant. English doesn’t have middles at all. The fact that Dr. Palmer used multiple languages should have been evidence to you that he was taking a much more typological approach.

  9. @Mike

    I’m arguing against describing Middle in terms that are familiar to a naive English, or French, or Proto-Indo-European speaker. Typology can come later, once a student has learned some Greek. But saying “it’s kinda like French or Spanish reflexives” from the get-go serves to misdirect the learner’s concept into a pre-existing category.

    To more seriously address your comment on voice: Van Valin includes a whole chapter on semantic vs. syntactic roles in his Introduction to Syntax (which has a survey of several theories, Chomskyan and non):

    “Thus, grammatical relations like subject and direct object are independent of semantic roles like agent and patient: in both English and Malagasy, the subject of a sentence can be the agent (active voice) or the patient (passive voice)…” (p. 23)

    Presumably a Chomskyan would avoid using an introductory textbook that used words like “agent” or “patient” in the same sentence as “subject” and “object”. Funny, I don’t see people complaining in the reviews, though…

    BTW, the fundamental principle of Van Valin’s Role & Reference Grammar is all about the relationship of the privileged syntactic argument to its semantic role, so it belongs in your second category: “Languages with voice constructions allow the default choice to be overridden: a passive construction permits the undergoer to be the privileged syntactic argument, whereas in an antipassive construction the actor is the privileged syntactic argument.” (p214)

Comments are closed.