[B-Greek] A new stab at voice

Carl W. Conrad cwconrad at artsci.wustl.edu
Tue Oct 16 12:14:55 EDT 2007

On Oct 16, 2007, at 12:39 AM, Kimmo Huovila wrote:

> I've been reading on voice, and am not very satisfied with most  
> that I have
> read. I realize that it would take quite a bit of research to  
> settle the
> matter. I know that there are others on the list that have done  
> more studies.
> However, I sketched a few preliminary thoughts on voice for  
> discussion, proof,
> or refutation, in hopes that it helps me get at the point of the  
> matter. Any
> comments?

Oh, yes! First of all, I'm delighted that somebody else has taken on  
a serious consideration of these matters. It was you, I think, Kimmo,  
who first put me to reading Kemmer, and after that Cooper.

Suzanne Kemmer's major study, _The Middle Voice_ (Amsterdam/ 
Philadelphia: John Benjamins Publishing Company, 1993).

One of the more useful recent formulations of the facts regarding  
voice usage in ancient Greek is the third edition of Rijksbaron:
Albert Rijksbaron, _The Syntax and Semantics of the Verb in Classical  
Greek_ (Chicago and London: University of Chicago Press, 2002).

With respect to ancient Greek since then, the most significant input  
has come, in my opinion, in the following works:

Egbert J. Bakker, “Voice, Aspect and Aktionsart: Middle and Passive  
in Ancient Greek” in Barbara A. Fox, Paul J. Hopper, edd., Voice:  
Form and function (Typological Studies in Language 27) (Amsterdam/ 
Philadelphia: John Benjamins Publishing Company, 1994), pp. 23-47.

Neva F. Miller, A Theory of Deponent Verbs (App 2 of Friberg,  
ANLEXGNT {Baker Books, 2000}, pp. 423-430).
Allen, Sydney J., The Middle Voice in Ancient Greek: A Study in  
Polysemy. Amsterdam: J.C. Gieben, Publisher, 2003. (Amsterdam Studies  
in Classical Philology, 11).

Allen, Sydney J., The Middle Voice in Ancient Greek: A Study in  
Polysemy. Amsterdam: J.C. Gieben, Publisher, 2003. (Amsterdam Studies  
in Classical Philology, 11).

Allen's study is thorough and deals with middle forms (including the - 
θη- {-QH-} "passive" forms) and usage in Homeric and Classical Attic  
Greek. He understands the semantic passive as a distinct category of  
middle usage closely akin to several other categories using the -θη-  
{-QH-} "passive" forms in the aorist, including Spontaneous process  
(e.g. ἐτάκην, ETAKHN, "melted), Mental process (e.g.  
ἐφοβήθην, EFOBHQHN, "feared"), Body motion (e.g.  
ἀπηλλάχθην, APHLLACQHN, "left behind"), and Collective  
motion (e.g. ἠγέρθην, HGERQHN (AGEIRW), "gathered"), as  
distinguished from categories of middle usage using the -μην/σο/ 
το (-MHN/SO/TO) "middle" forms in the aorist, including Reciprocal  
(ἐμαχεσάμην, EMACESAMHN, "fought"), Direct Reflexive  
(e.g., ἐλουσάμην, "washed"), Perception (e.g.  
ἐγευσάμην, EGEUSAMHN, "tasted"), Mental activity (e.g.  
ἐλογισάμην, ELOGISAMHN, "reckoned"), Speech act  
(ἐλοιδορησάμην, ELOIDORHSAMHN, "reviled"), and Indirect  
Reflexive (e.g. ἐποιησάμην, EPOIHSAMHN, "performed for  
self"). Allen notes that the -θη- {-QH-} "passive" forms gradually  
grow more frequent in categories in which the -μην/σο/το (-MHN/ 
SO/TO) "middle" aorist forms were predominant earlier -- and this is  
something that continues in Hellenistic and later Greek, where we  
must understand both "middle" and "passive" usages as represented by  
the same -θη- {-QH-} "passive" forms.

Allen's work needs to be extended into the realm of Hellenistic  
morphology and syntax. Most of my own researches have been in the  
area of NT and LXX forms and usage, but a thoroughgoing exposition of  
Biblical Greek forms and usage along the lines of Allen's work is  
very much in order. There's really very little that's been done in  
recent years: Jonathan Pennington has discussed the matter of  
deponency in Biblical Greek verbs, an SBL paper “Is Deponency a  
Valid Category for Koine Greek?” in November 2003, and a paper,  
"Deponency in Koine Greek: The Grammatical Question and the  
Lexicographical Dilemma" published inTrinity Journal 24 (NS) (Spring,  
2003), 55-76, and Bernard Taylor's paper, "Deponency and Greek  
Lexicography" in _Biblical Greek Language and Lexicography_, ed.  
Taylor, Lee, Burton, Whitaker, Eerdmans, 2004, pp. 167-176, all  
question continued use of the term "deponent" to refer to such verbs  
as others have more accurately termed "media tantum" and "passiva  
tantum" with respect to their standard aorist morphology.

I am pretty much in accord with what Kimmo has expressed here, and my  
own formulation of many of the same propositions may be found on my  
web-page, "Ancient Greek Voice" (http://www.ioa.com/~cwconrad/ 
GrkVc.html). I'll comment on some of these items below.

> 1) Transitivity is unrelated to voice and should not be confused  
> with it.
> There is nothing inherently strange in an intransitive active verb  
> or a
> transitive passive or middle verb.

"Subject-affectedness" rather than transitivity is the chief feature  
distinguishing the voice-forms in ancient Greek. Verbs with "active"  
morphology are UNMARKED, while verbs with "middle" or "passive"  
morphology are MARKED for subject-affectedness. But transitivity is a  
factor in the differentiation between "middle" and "active" forms of  
quite a few verbs: there is a sizable category of middle-voice verbs  
of Spontaneous process, Mental process, Body motion, and Collective  
motion that are intransitive but that have active-voice causative  
forms that are transitive (e.g. hISTAMAI/hISTHMI, AGEIROMAI/AGEIRW,  
FOBEOMAI/FOBEW). Egbert Bakker's article cited above nicely sketches  
out the interrelationship between voice, transitivity, and causality.

> 2) There is nothing inherently strange for a verb to lack any of  
> the voices.
> (It is admitted that some verbs regularly have more than 1 voice,  
> for example
> transitive actives, which can very often be passivized. The point  
> here is
> rather that it is not strange that there are SOME verbs with no  
> active,
> middle, or passive. Naturally the verb frame and semantics matter.)

Another factor in this regard is that some everyday verbs have  
aorists and perfects which hark back to forms from an era when there  
was no distinct middle-passive alternative forms, e.g. hESTHKA (orig.  

> 3) The subject of the active voice can be in almost any semantic  
> role of the
> verb frame. However, it is not lower in the following hierarchy  
> than the rest
> of the roles of the verb frame (agent>author>instrument>patient).  
> Note that
> PASCW hUPO is not a counter-example, as the subject is an  
> experiencer. This
> rule could probably be elaborated to cover more ground.

I think this is best expressed by saying that the active-voice forms  
are UNMARKED while the middle-voice forms are MARKED for "subject- 
affectedness." Subject-affectedness may characterize active-voice  
verbs but in such verbs it isn't marked.

> 4) The subject of a middle fulfills two roles in the sentence. The  
> other may
> be a beneficiary, in which case the semantic load of the middle (as  
> opposed
> to the active) may be small. The middle may be reflexive or  
> reciprocal.

i.e. the subject may be a receiver or a patient.

> 5) The subject of a passive sentence can be (depending on the verb)  
> in almost
> any other semantic role than the one usually filled by the subject  
> in the
> corresponding active sentence. It may represent the accusative,  
> genitive,
> dative or a prepositional phrase in the active if the verb is also  
> used in the
> active.


> 6) The subject of a passive verb is not the agent. It does not  
> follow that the
> subject is a passive participant. The subject may have indirect  
> control over
> the event. POREUOMAI is construed as the subject being the patient of
> conveying (done for example by horse). It could be construed as an  
> agent
> (as in some languages like English 'travel'), but the Greek  
> language happens
> to construe the semantics differently. The patient in this case is a
> participant, which usually has caused that he is conveyed, but he  
> is not
> the agent of conveying (as a donkey for example would be the  
> agent). The verb
> could be semantically middle, but it just happens to be passive.  
> in the passive does not mean that the subject of baptism is passive  
> in it. He
> may get himself baptized, but the expression does not focus on his  
> role as
> being active, but as receiving baptism. There is no need to  
> interpret either
> of these verbs as semantically middle (though a middle could do as  
> well).
> There is enough overlap in the situations that can be expressed  
> with a middle
> and a passive.

It would, I think, be better to speak of "agentive" and "non- 
agentive" verbs and to note whether the action/process indicated by  
the verb is voluntary or involuntary (spontaneous). I would  
personally say that EBAPTISQHSAN ought not to be understood as a  
"passive verb" in the sense that the subject is solely the patient;  
rather I agree with Kimmo that willing participation in the process  
is clearly involved here. The difference between the form EBAPTISANTO  
and EBAPTISQHSAN lies in the fact that the former indicates the  
subjects performs the action of baptism upon persons other than  
themselves, whereas the latter indicates that the subjects have  
themselves undergone the action of baptism, presumably voluntarily  
(for if not, what validity would the baptism have?).

> 7) The substantial overlap in semantics contributes to the  
> distinction not
> being strict. Sometimes the passive morphoparadigm is semantically  
> middle.
> EGERQH- is often most naturally interpreted as having the subject  
> as agent
> and patient. This is common middle semantics. "TI EPOIHSAS? -  
> HGERQH." should
> not be strange, as it would be if HGERQH was semantically passive.  
> Therefore
> the morphological distinction between the passive and the middle is  
> not
> complete.

I agree.

> 8) It does not follow from 7 that the distinction is not real or  
> that there is
> no semantic difference between the morphoparadigms in some cases.  
> Proof for
> the distinction is asymmetry: the passive morphoparadigm can be a  
> passive of
> either the active or the middle morphoparadigm, but the middle  
> morphoparadigm
> cannot be a passive for the passive morphoparadigm.

I quite agree. Bakker (cited above) rightly notes the importance of  
Aktionsart and transitivity in relationship to the semantic force of  
the morphoparadigm representing the action/process in particular  
instances. (By the way, I'm happy to see you're employing my coinage,  

> Any thoughts?

Thank you, Kimmo, for a stimulating report. Keep at it!

Carl W. Conrad
Department of Classics, Washington University (Ret)

More information about the B-Greek mailing list