MIDDLE AND PASSIVE VOICE
Carl W. Conrad
cwconrad at artsci.wustl.edu
Thu Oct 25 10:52:26 EDT 2001
At 4:12 PM +1000 10/25/01, B. Ward Powers wrote:
>This is a response to the important posting from Carl about voice in Greek.
>Actually, the gap between our positions is not as great as may at first
>appear. But it is certainly there. It is a difference which to an extent
>lies in how we describe the observed phenomena of the language; but it does
>have some implications for meaning and interpretation.
Let me first endeavor to spell out what I see as points on which we agree
and points on which we differ:
We agree: that quite a few very important NT Greek verbs have forms with
MAI/SAI/TAI and -QH- morphology that are intransitive or active in sense.
Ward uses the conventional designations for such forms: "middle deponent"
(verbs that have 'middle' morphology in the aorist and future tenses) and
"passive deponent" (verbs that have 'passive' morphology in the aorist and
We disagree on two significant matters, and to different extents; if I am
misrepresenting Ward's view, I hope he will correct me:
(1) Ward supports the conventional understanding that -QH- forms in the
aorist and future tenses signals an authentic passive meaning EXCEPT in the
case of the so-called "passive deponents"; MY working hypothesis (and I
won't claim that it is fully demonstrable to my own satisfaction as yet) is
RATHER that -QH- forms in the aorist and future tenses did, in the course
of the historical growth and development of ancient Greek, gradually and to
a considerable extent SUPPLANT older aorist 'middle-passive' morphology and
then, in those verbs thus supplanted, also formed new future tenses in
-QHSOMAI to supplant the older 'middle-passive' -SOMAI morphology.
(2) While admitting that the term "deponent" is "just a handle for a
phenomenon; nothing more," Ward uses the term in precisely the fashion
which for my part, I find disturbing for two reasons: (a) it involves a
confusion of transitivity with voice morphology; and (b) it misleadingly
suggests that there is something deficient in these verbs, whether as
consequence of loss by these verbs of an 'active' morphology or by failure
to observe/preserve the proper relationship between the voice morphology
and the voice 'concept.' Let me draw upon Ward's latest message to show
what I mean:
At 4:12 PM +1000 10/25/01, B. Ward Powers wrote:
>It will be noted that I have said, re this word "deponent", "The term is
>just a handle for a phenomenon; nothing more."
>Carl questions "the entire concept of deponency", categorizing it as "worse
>than useless", adding, "Nor is it valid or helpful to define, a 'deponent'
>verb as one that has middle or passive form but 'active' meaning."
>This pinpoints a place where we differ. Morphologically, certain morphs
>(especially the endings) are associated with the concept "middle", and
>other morphs (the presence of -QH/QE-) with the concept "passive". We
>should not lose this correlation, nor play it down: it is valuable.
>But then we note that some verbs (e.g., DUNATAI) are middle in form, and
>others (e.g., APOKRIQEIS) are passive in form, while the meaning is active.
>We want to be able to discuss this language feature. I am happy to use the
>word "deponent" because it is an existing and quite widely-used term which
>can therefore be a useful starting point for discussing this language
>feature. But I am not irrevocably wedded to the use of this term, and if
>Carl has one which can win acceptance instead, I have no objection to it.
>But we need SOME term to use when we want to talk about the fact that
>certain verbs have a form which correlates with normal middle use and/or
>normal passive use, while not themselves having middle or passive
>signification in their context.
I find two points in what Ward says here to be problematic:
(a) First, there is the linkage of MP morphology and -QH- morphology with
the CONCEPTS "middle" and "passive. I seriously doubt that there is any
distinct CONCEPT "passive" intrinsically or necessarily linked to the -QH-
morphology; I believe rather that what I call the "subject-intensive"
semantic character associated with the MP morphology lends itself, when
needed, to express a PASSIVE semantic conception: that BALLETAI although
essentially middle, can and does quite regularly convey the passive sense,
"is thrown/put." Such a verb as APOLLUTAI can assume either a middle sense,
"goes to waste" or a passive sense, "is being ruined/destroyed." I really
think that the same is true of the -QH- forms in the aorist and future:
that they can assume EITHER a middle OR a passive sense, and that they are
NOT intrinsically and necessarily linked to the passive CONCEPT.
(b) Secondly, although I don't know whether Ward intended to assert that
the verb DUNATAI is active in meaning--I would certainly say that it's
intransitive and normally completed by an infinitive--, but when "deponent"
is defined, it seems that what pops into the definer's head at once is:
"it's a verb that's middle or passive in form but active in meaning." I
don't believe there's any intrinsic or necessary link between transitivity
and the morphological paradigms that we conventionally term "active,"
"middle-passive" and "passive." I think that the conventionally-termed
"active" morphology is employed in verbs that are semantically active (e.g.
LAMBANEI), semantically intransitive (e.g. BAINEI) and even semantically
passive (e.g. PASCEI--note that the accusative object of PASCEI refers to
what is done TO the subject); and I think that the conventionally-termed
"middle" morphology is employed in verbs conceived by the Greek psyche as
"subject-intensive" whether the verb is 'active' insofar as it takes a
direct object (e.g. DECETAI, or intransitive (e.g. POREUETAI) or
semantically 'passive' (e.g. BALLETAI, "is thrown/put").
Do I have an alternative suggestion for a word to take the place of
"deponent"? No, I don't, and I really don't want to replace it but to
scuttle it. I believe that the term derives from an inadequate conception
of the nature of voice and of the relationship between voice morphology and
voice semantics. I think we need to learn verbs like DUNAMAI and POREUOMAI
and APOKRINOMAI as "subject-intensive" verb forms and list their principle
parts as they actually occur along with their meanings:
I should add that I also believe a re-naming of the morphological
paradigms/endings is also in order, in accordance with what I would see as
a "paradigm shift with regard to Greek voice." What I suggest is not very
good, but at least it does not suggest misleading notions: I would re-name
what is conventionally termed "active" morphology as "default" or perhaps
"normal" or even "standard"--certainly most verbs in Greek are conjugated
in the W/EIS/EI paradigm; secondly, I would re-name what is conventionally
termed "middle-passive" (but is all too often simply called "middle")
morphology as "subject-intensive." And finally I would re-name the -QH-
paradigms of the aorist and future ALSO as "subject-intensive." As I noted
yesterday, I don't expect such a "paradigm shift with regard to Greek
voice" to take place sooner than the United States adopts the metric
system. If either of these ever happens, I'm sure it won't be in my
lifetime (but what do I know? I never thought the Berlin wall would come
down or South Africa become democratic in my lifetime!).
In my earlier message to which Ward was responding I had written:
>>I'm going to comment only on a very few items in Ward's presentation
>>because I honestly believe more of it concerns explanation of exceptions to
>>the rule than clarification of the rules themselves. I don't mean to be
>>unfair, but that is my perception.
>I would NEVER consider Carl to be unfair. But let me clarify the issues. My
>approach homes in on relatively simple "rules" which cover the majority of
>cases. Because of their relative simplicity, and (I would believe) clarity,
>not much needs to be said about them....
Although I certainly did NOT mean to be unfair in my characterization of
Ward's previous message, I think I was, in fact, unfair. His point is quite
valid: insofar as the rules are clear and simple, they don't require a
lengthy statement; it is the exceptions, such as are to be found to just
about any rules set forth regarding grammatical patterns, that will
necessarily require lengthier exposition. And so Ward says, quite rightly:
>> >Every known language has language rules - and exceptions to these rules.
>> >The existence of the exceptions does not nullify the rules. The rules
>> >provide the basic pattern for most forms in that language, and the
>> >foundation on the basis of which you can identify, and handle, the
>I still stick to that view.
But are there possibly so many exceptions that the rules themselves should
be called into question? What started me thinking along the lines in the
first place was recognition that HGERQH in the GNT far more often than not
is NOT semantically passive but identical with ANESTH and intransitive
semantically: "he rose"--NOT (at leasat not ordinarily) "he was raised." I
started looking at more and more verbs with the -QH- morphology and with
the -MAI/SAI/TAI morphology that were commonly parsed as "passive" and
discovered that many were CLEARLY "subject-intensive" semantically while
some were indeed CLEARLY passive semantically, while others still were
When I started exploring what the historians of Greek grammar, Meillet,
Chantraine, Sihler, as well as what A.T.Robertson and others have written
regarding the development of the -QHN forms as an extension of the
essentially stative/intransitive -HN "3rd aorist" morphology and the
gradual supplanting of the MAI/SAI/TAI forms in the aorist and future by
these, I started the investigation of actual forms and usage in the GNT
that has brought me to the present stage of my thinking and formulation of
my hypothesis regarding voice. I am now in the process of looking at GNT
verbs individually and finding things that have surprised me. One
conviction has stuck fast with me, however: the Greek language was in flux
during the NT era and concurrent older and newer forms of the same
constructions and morphological structures were simultaneously in use; the
synchronic approach to NT Greek grammar, I am quite convinced, is not
adequate to elucidate the "rules" and "exceptions" of Koine usage. Snippets
of information are not enough to validate hypotheses, but awareness of such
facts as that Homeric BLHTO meant "was thrown" and that the conjugation of
EINAI in Modern Greek is wholly Middle are suggestive.
In my post to which Ward was replying, I said:
>>I have done only a preliminary tally of the forms of LUW in the GNT and
>>find that there are 42 hits; of these 15 are tagged as "passive" in
>>Accordance, not one as middle. The seven forms in the aorist are all -QH-
>>forms: 2 indicatives, 1 subjunctive, 3 infinitives, 1 participle. The three
>>forms in the future are all -QH- forms and are all indicative. So we may
>>well TEACH a paradigm including LUSOMAI and ELUSAMHN, but in fact these
>>forms don't appear in NT Koine Greek.
And Ward has replied:
>I am not in a position to comment on Carl's figures, but I have no basis
>for disagreeing with him. But this is not the point. The Greek verb is
>usually taught (certainly in my Grammar "Learn To Read the Greek New
>Testament" and in other Grammars with which I am familiar) using LUW as the
>pattern verb. This is because (being a long-vowel verb, and a very short
>word) it is ideally suited for this purpose, irrespective of its meaning,
>its NT frequency, or its occurrence pattern. I seem to recall reading an
>email from Carl in which he mentions that LUW was the pattern-verb used
>when he also learnt Greek. (Have I remembered this correctly, Carl?) The
>forms of LUW which Carl points out do not occur in the GNT actually DO
>exist in the language and DO occur for other verbs for which LUW has been
>our pattern to learn them.
>So the morphological meanings which we derive from using LUW as a pattern
It is true that in my April 1999 discussion of Greek Voice I used middle
forms of LUW to illustrate how the passive meaning, while not intrinsic to
what I would now term the middle's "subject-intensive" sense, comes to bear
a passive sense when an instrumental dative or a hUPO + genitive agent
construction is employed with it. At the time I wasn't even aware that the
"middle" LUW does not appear in the GNT EXCEPT where it bears a passive
sense, and that the only aorist and future forms of it in the GNT are -QH-
SHOULD the aorist ELUSAMHN and future LUSOMAI paradigms be taught in a
Koine NT Greek course? Well, yes, if one's going to learn the Optative
forms also, although one will meet very few of them while reading the GNT.
Certainly one will meet futures in -SOMAI for several important traditional
verbs such as MANQANW, LAMBANW, BAINW, but I really do wonder how many
Greek verbs in the GNT actually do have BOTH 'middle' aorist in -SAMHN and
future in -SOMAI AND 'passive' aorist in -QHN and future in -QHSOMAI. I'm
going to try to get the figures to resolve this question before I'm
through. I suppose the question at root is what are the morphological
boundaries of the language one intends to present in strictly SYNCHRONIC
terms as "NT Koine Greek"?
Finally, let me comment on Ward's closing remark, because in it he affirms
a proposition that I firmly espouse as strongly as does he:
>Greek is a very morphologically-controlled language. If you can recognize
>and understand the morphs, you will almost always get a clear indication of
>meaning. The relatively rare exceptions occur where context and/or usage of
>a particular verb overrule the normal significance of the morphs; and we
>can note these when they occur without losing the benefit of the general
>rules. In my previous b-greek posting I said.
And in his earlier message, Ward said:
>> >There is no "one right way" of describing what is found in a language - all
>> >we can aim for is what you might call the "best fit" of a description. But
>> >I definitely do not believe that Carl's approach to middle and passive is
>> >the best way of approaching what we find in Greek, and teaching it.
I quite agree with both these formulations--and yet each of us perceives
the phenomena of the Greek voice system (and of the pedagogical needs) from
very different perspectives. I don't disagree at all with the proposition
that "Greek is a very morphologically-controlled language. If you can
recognize and understand the morphs, you will almost always get a clear
meaning." The single greatest point in dispute between us is that I don't
really believe that the -QH- morphs have a distinct and necessary linkage
to the passive "concept", and I strongly suspect that the sizable aggregate
of so-called "middle deponents" and "passive deponents" may more properly
be said to conform to the "rule" rather than represent "exceptions to the
Once again, my compliments to Ward, both for being clear and precise in the
formulation of his understanding of these matters and no less for providing
me an opportunity to clarify in my own mind and find appropriate expression
for what I am groping toward--not so much a radical new discernment of the
linguistic phenomena as a new way of understanding the patterns into which
those phenomena fall.
Carl W. Conrad
Department of Classics, Washington University (Emeritus)
Most months: 1647 Grindstaff Road/Burnsville, NC 28714/(828) 675-4243
cwconrad at artsci.wustl.edu OR cwconrad at ioa.com
More information about the B-Greek