MIDDLE AND PASSIVE VOICE
Carl W. Conrad
cwconrad at artsci.wustl.edu
Sat Oct 27 13:36:51 EDT 2001
I'm responding here to two messages from Iver; where the two messages
overlap (esp. regarding interpretation of forms of EPISTREFW), I respond to
the formulation in the second message. I'm going to try to eliminate
citation of previous correspondence to what is essential to what I want to
At 12:06 AM +0200 10/27/01, Iver Larsen wrote:
>The discussion about voice has been interesting to me, but it seems to be
>quite complex. Let me add some comments from my perspective. I am not
>"taking sides" just looking at the issue from a different angle.
For my part, I have found these comments very helpful in three regards: (1)
clarification of the understanding of transitivity, (2) the question of
appropriate names for the "middle-passive" (MAI/SAI/TAI, etc.) and
"passive" (-QH-) morphological paradigms, (3) the "Eigenart" or
'distinctive quality' of the Greek 'middle' voice, and (4) implications of
the possibility that -QH- paradigms may be 'middle' or 'passive'.
(1) clarification of the understanding of transitivity
>In order to hopefully avoid confusion, let me just briefly state
>the >terms I am using:
>Intransitive, transitive and ditransitive verbs are syntactic terms and
>describe whether a verb when it has its complete set of arguments is
>constructed with just a subject, a subject and an object or a subject and
>two objects, one of which may be indirect object.
>The term passive I understand as a derived form of a transitive or
>ditransitive verb form. This means that an intransitive verb cannot have
>In semantics, it is more common to talk about verbs that have one, two or
>three basic valencies. These valencies are like arms that go out from the
>verb nucleus and are able to grab one to three nominals in various roles.
>The primary or basic semantic roles are: 1) agent (with subclasses:
>experiencer and cause), 2) patient or undergoer, 3) beneficiary or location.
>The syntactical passive corresponds to making the agent role implicit.
>This means that either the patient or beneficiary becomes the grammatical
This is very helpful. The one point about which I am not yet clear and
comfortable is the definition of 'passive' as a "derived" form, but it may
be that I am thinking too specifically of Greek and what seems to me to be
distinctive about how the Greek middle-passive paradigms (both MAI/SAI/TAI
and -QH-) actually function to focus on the subject as patient/undergoer or
beneficiary. It seems to me that both the M-P paradigms focus on the
subject without even implicitly entailing an agent or cause, and that these
paradigms BECOME passive when and only when an agent or cause becomes
clearly implicit or explicit. But I may be confounding semantics with
(2) the question of appropriate names for the "middle-passive"
(MAI/SAI/TAI, etc.) and "passive" (-QH-) morphological paradigm
>I agree with Carl that the word "active" is not the best term because one
>can be tempted to equate "active" with "transitive". This does not
>hold >for Greek since several intransitive verbs are "active" in
>their >morphology. Nor do I think it is helpful to connect "middle"
>morphology >too closely with "intransitive verbs". Many middle forms are
>transitive >verbs. Carl has suggested new names for the "active" paradigm
>such as >default, normal or standard. I would prefer "basic", because I
>would >still like to see the passive as derived from a basic form. Carl
>has >stated that there are a number of verbs in Greek with
>"passive" >morphology that are not passive in meaning. If I understand
>him >correctly, he prefers to lump the middle and passive together into
>one >group which he calls "subject-intensive". I am not too happy with
>this >term nor with lumping the two into one. I think the vast majority
>of the morphologically passive forms are genuinely passive in that
>they are derived from a basic form of the same verb. On the other
>hand, >it seems that some passive aorist or future forms could well
>middles, and a few even as active. I am not sure a form like BALLETAI is
>middle in sense. It seems to be passive in KOINH.
We are agreed that the W/EIS/EI (including MI/S/SI) paradigms really ought
not to be called "active." What to call them remains a problem. I really
have no objection to "basic" BUT I may not be referring to the same thing
as Iver refers to, since I don't see any derivation of the "passive" (-QH-)
morphology from the "basic" (i.e. 'active') forms of the verb root or
tense-stem at all. Historically it seems to me derivative from verb root
and stem of the oldest non-thematic aorists in HN/HS/H (e.g. EBHN/EBHS/EBH,
ESTHN/ESTHS/ESTH) in the same way that future stems are often derivative
from aorist stems rather than from present stems (e.g. TEX-OMAI from
ETEKON, not from TIKTW, GENHSOMAI from E-GEN-OMHN, not from GI(G)NOMAI).
But as for a proper term to designate what has traditionally been called
'middle-passive' (i.e. the MAI/SAI/TAI paradigms), I'm not wedded to
"subject-intensive" and upon lengthy reflection I can see that 'intensive'
is indeed misleading: these forms don't express anything of what an AUTOS
coupled with the implicit or explicit subject of a verb expresses. Now I'm
wondering how "subject-focused" might fit as a technical term of reference
for these paradigms: what they involve is certainly a focus on the subject
as patient, experiencer, undergoer whether-or-not he/she is also an agent.
And that leads directly into the next thing I want to respond to in your
(3) the "Eigenart" or 'distinctive quality' of the Greek 'middle' voice
>The "middle" forms are the most tricky to understand and describe.
>It >would be useful to make a study of all the verbs that have both
>"active" >"middle" and "passive" forms to try to see a general pattern, if
>there is >one. In >KOINH Greek there seem to be few verbs with a middle as
>well as >passive paradigm in either future or aorist. So, is there a
>standard >description of the voice differences in Classical Greek, one
>that does >not equate "active" >with transitive and "middle" with
As I mentioned a couple days ago, when Rod Decker, in response to my
listing of Middle Categories from Suzanne Kemmer's dissertation on the
middle voice, mentioned the essay, I rushed at once to order a copy from
amazon.com of the Fribergs' _Analytical Lexicon of the Greek New Testament_
for the one and only reason of reading Neva Miller's short essay entitled
"A Theory of Deponent Verbs" in the appendix. As for the book itself, it
seems to me that in order to use it as a parser pony the student will need
to learn an arcane tagging code sufficiently complex that his efforts would
better be repaid by learning Greek instead. And yet this essay I deem worth
the price of the book--I only regret learning that the author died three
years ago and is no longer with us and accessible to discuss these matters.
Here's what she says (p. 426):
"... the meaning in the verb involves significant movement that
comes back in some way to cause the agent of the action also to become
affected by that action. In other words, an emphasis is put on reflexive
action, and the subject, when he is the agent of the action, becomes the
center of gravity. The agent does something that benefits himself. The
action is not transferred away from him, since the action in the verb does
not pass through to affect an object that is only outside of him. He stays
involved. for example, in the verb fight, the action in the verb is
meaningless unless the subject stays involved in that action (recall the
saying "It takes two to fight"). It is hard to imagine what the original
active form, if such existed, would have had as its meaning for verbs like
answer, try, doubt, fear, touch.
"I propose, then, that we need to work toward a better
understanding of those verbs for which no active form is found We need to
examine each such verb for its own sake and allow it to speak for itself.
Since the middle voice signals that the agent is in some way staying
involved in the action, it is appropriate to ask, How is the agent
involved? Is he benefiting himself (e.g., I eat)? Is he interacting
positively with someone else (I welcome)? Or is he interacting negatively
with someone else (I leap on)? Could it be that he is communicating with
someone else, so that if he did not stay involved as the speaker, the verb
would become meaningless? And how could a person feel ashamed unless there
were interaction with his own thoughts and feelings?"
She then goes on to offer her own list (what Rod referred to in response to
my citation of Kemmer's list) of types of 'middle-voice' or 'deponent'
verbs. Here's her list:
Class 1: Reciprocity
A. Positive [i.e. friendly] Interaction
B. Negative [i.e. hostile] Interaction
C. Positive and Negative Communication
Class 2: Reflexivity
Class 3: Self-Involvement
A. Intellectual Activities
B. Emotional States
C. Volitional activities
Class 4: Self-Interest
Class 5: Receptivity
Class 6: Passivity
Class 7: State, Condition
Into these categories, which are (as Rod Decker noted) conceived and
ordered in a different way from that developed by Suzanne Kemmer, Miller
has arranged just about all, if not absolutely all, the so-called
'deponent' verbs found in the GNT. Whether or not one finds this scheme
fully satisfactory, it is instructive, I think, for our understanding of
what these verbs are and what they have in common.
(4) implications of the possibility that -QH- paradigms may be 'middle' or
Before commenting on EPISTREFW, I want to just note that I am now in the
process of examining individually all GNT verbs that have either/and/or
aorist forms in the 'middle-passive' (MHN/SO/TO) paradigm or in the
'passive' (-QH-) paradigm. One question I have been seeking clarification
about is whether Ward Powers is right in asserting that the differentiation
between the aorist middle morphology and the aorist passive morphology is
really significant. Even statistically, what I've found is (to me, at
(a) There are 1328 aorist 'middle' verb-forms in the GNT representing 194
verbs that have aorist 'middle' (MHN/SO/TO) paradigms;
(b) There are 1775 aorist 'passive' verb-forms in the GNT representing 340
verbs with aorist 'passive' (-QH-) paradigms;
(c) There are only EIGHTEEN (18) verbs represented by forms in BOTH
'middle' and 'passive' paradigms. These are:
AGALLIAW (4x aor. 'middle', 1x aor. 'passive', identical sense)
ANAPAUW (12x aor. 'middle', 1x aor. 'passive', identical sense)
ANATREFW (1x aor. 'middle', 1x aor. 'passive'
APOKRINOMAI (7x aor. 'middle', 213x aor. 'passive'
BAPTIZW (2x aor. 'middle', 32x aor. 'passive'
DIAMERIZW (3x aor. 'middle', 1x aor. 'passive'
DIATASSW (3x aor. 'middle', 3x aor. 'passive'
EKTIQHMI (1x aor. 'middle', 1x aor. 'passive'
ENDUW (19x aor. 'middle', 6x aor. 'passive'
EPIKALEW (7x aor. 'middle', 11x aor. 'passive'
EUAGGELIZW (18x aor. 'middle', 4x aor. 'passive'
hEURISKW (1x aor. 'middle', 27x aor. 'passive'
KATALAMBANW (3x aor. 'middle', 1x aor. 'passive'
LUTROW (1x aor. 'middle', 1x aor. 'passive'
MERIZW (1x aor. 'middle', 1x aor. 'passive'
METAPEMPW (7x aor. 'middle', 1x aor. 'passive'
hORAW (root OP) (1x aor. 'middle', 23x aor. 'passive" (= "appear")
Isn't it interesting that ONLY 18/534 verbs in the GNT with aorists in
'middle' and/or 'passive' paradigms have forms in BOTH paradigms? I think
this MUST be interpreted to mean that the -QH- paradigms have already
become stabilized to represent the "subject-intensive" or "subject-focused"
counterpart to the "basic" ('active') paradigms. And I think that it means
we have concurrent forms in the period of GNT KOINH Greek (I think it's
really not very different from the concurrent forms of the "first" or -SA
aorist and the "second" or -ON/ES/E aorist).
Granted, a great many of the verb-forms in -QH- bear a real PASSIVE sense,
nevertheless, the whole slew of so-called "passive deponents" (aorists in
-QH- that aren't passive in meaning) as well as numerous -QH- verb forms
that are NOT "passive deponents" but that can be shown to have a "middle"
or intransitive sense seem(s) to indicate that the -QH- paradigms are NOT
distinctly and fundamentally passive in function and meaning.
Now, to the forms of EPISTREFW;
>Since Kimmo mentioned EPISTRAFEIS in John 21:20 I looked at that.
>There >are 5 aorist "passive" examples of this verb in the GNT, no middle
>forms >and a good number of basic forms. Some of the basic forms are
>transitive >in the sense that the agent is different from the patient, e.g
>James 5:20 hO EPISTREYAS hAMARTWLON he who has turned a sinner back (also
>But most seem to have the agent and patient combined so that the
>agent >turns himself. I am not sure if this is what Carl
>calls >"subject-intensive", but with this verb the active forms are used,
>not >the middle. When looking at the active as opposed to middle-passive
>it >seems to me that the MP forms suggest a turning around oneself or
>a >passive sense of being returned to a former position by another
>agent. >The active forms suggest a change of direction, a turning of
>oneself, but >turning towards some other person or in a certain direction,
>not just >turning around on the spot. I cannot list all the examples, just
>a few, >first the active forms:
>Mt 12:44 EIS TON OIKON MOU EPISTREYW to my house
>Mt 13:15 EPISTREYWSIN - they might turn to me (to me is implied)
>Mt 24:18 MH EPISTREYATW OPISW - let him not turn behind - back to his house
>Lk 17:4 EPISTREYHi PROS SE - if he turns to you
>Acts 9:35 hOITINES EPESTREYAN EPI TON KURION -who turned to the Lord
>Acts 15:36 EPISTREYANTES DH EPISKEYWMEQA - having turned to where we came
>from, i.e. returned.
>Rev 1:12 EPESTREYA BLEPEIN THN FWNHN - I turned (towards the one speaking)
For my part I would NOT call any of these 7 forms of the verb
"subject-intensive" and certainly not active (as unquestionably James 5:19
and 5:20 are active); rather I would say ALL of them are INTRANSITIVE;
EPISTREFW is a verb of motion just like BAINW. There's the directional
prefix EPI- on EPISTREFW but even if there is personal physical exertion
here in the turning around in a direction, so also is there in BAINW with
its moving the feet apart from each other. I think the sense of the verb
certainly lends itself to what I understand the sense of the 'middle' to
be, but in fact, even in the future, it remains 'active' or 'basic' in
voice, while on the other hand, BAINW becomes BHSOMAI as do so many verbs
that are 'active' or 'basic' in the present tense.
>Then the MP forms:
I'm wondering, first of all, whether Iver really meant to call these
"middle-passive" forms; in traditional terminology they are all "passive"
forms, although for my part I would quite readily call them all
"subject-focused" (to use my latest coinage). In any case, what I find
gratifying is Iver's recognition that at least some of these forms are
surely NOT passive in meaning.
>Mt 10:13 hH EIRHNH hUMWN PROS hUMAS EPISTRAFHTW - let your peace
>be >returned to you
>Mk 5:30 EPISTRAFEIS EN TWi OCLWi ELEGEN - having turned around (himself)
>inside the crowd he said
>Mk 8:33 EPISTRAFEIS KAI IDWN TOUS MAQHTAS AUTOU - having turned around
>(himself) and having seen his disciples...
>Jn 21:20 EPISTRAFEIS hO PETROS BLEPEI - Peter having turned
>around >(himself) he sees
>1 Pe 2:25 ALLA EPESTRAFHTE NUN EPI TON POIMENA - but now you have been
>(re)turned to the shepherd
>My suggestion is that the verb EPISTREFW is semantically a divalent verb
>with two roles: An agent role and a directional role which indicates the
>direction towards which one is turning. Whether the verb is grammatically
>transitive or intransitive is not of particular interest to me.
>In the active forms above we have a full form with both roles. A
>few >times the second role is a patient, expressed in the grammar as
>object. >(That would make it grammatically transitive as in "I returned
>the >book"). But in most of the examples, the second role is not patient,
>but >a direction. (That makes it grammatically intransitive as in
>"I >returned".) Whether transitive or intransitive, it is still divalent
>with >two semantic roles. The direction role is shown in the grammar by
>a >location adverb (OPISW), by a preposition (EIS, PROS, EPI) or it
I think this is eminently sensible; I wouldn't quarrel with any of this.
>In the two MP forms that function as passive the patient role has been
>changed (or has changed? - the English middle?) from object to
>subject. >In Mt 10:13 peace as an abstract "object" has been returned to
>where it >came from. And in 1 Pe 2:25 the people have been turned back to
>where >they came from, the shepherd.
I don't agree that Mt 10:13 and 1 Pet 2:25 are passive; I think that both
are intransitive. Consider:
Mt 10:13 KAI EAN MEN Hi hH OIKIA AXIA, ELQATW hH EIRHNH hUMWN EP' AUTHN,
EAN DE MH Hi AXIA, hH EIRHNH hUMWN PROS hUMAS EPISTRAFHTW
If the 'coming' of 'your peace' upon the house is intransitive, then why
isn't the 'returning' as well? It would appear that this "Shalom!" is much
like a good check or a bad check: either it "goes through" or it "bounces."
And although the tellers are involved either way, all that the verb-form
shows us is the action itself, just as in hUEI "it's raining" or CALAZEI
1 Pet 2:25: HTE GAR hWS PROBATA PLANWMENOI, ALL EPESTRAFHTE NUN EPI TON
POIMENA KAI EPISKOPON TWN YUCWN hUMWN.
Here too it seems to me that the addresses, like the sheep, have of their
own accord wandered off-course, but now they have returned to the
shepherd--apparently deliberately, and I think that EPESTRAFHTE indicates
not an external agent or cause any more than does PLANWMENOI: I think that
both verbs are what I'm (for today, at least) calling "subject-focused"
>The three other MP forms that function as "middle" are not derived by
>suppressing the agent and moving the patient into subject slot, but by
>suppressing the direction. What they have in common is a change from
>a >divalent verb to a monovalent verb. When you say "Peter turned
>around" >you cannot add a directional role.It is complete with only one
>combined >role that is both the agent, experiencer and directional goal,
>expressed >in the grammar as subject.
And I'll reiterate that in my opinion all five of those verbs should be
deemed "middle" or "subject-focused."
>I don't think this explains everything about middle, but it is helpful
>for me to look at it from a semantic viewpoint.
>There seems to be an intensive concept in certain middle verbs, and my
>suspicion is that this is especially the case in classical Greek but has
>become rare in Koine. However, I do not think the majority of MP forms
>can be explained as intensive.
And I agree with and thank Iver for moving me away from the
"subject-intensive" designation. Nevertheless the subject is involved or in
focus in a way that is not the case with the "basic" or "active" paradigm.
Let me add as a footnote, for what it's worth, that one notion I've been
exploring is the possibility that Greek, like Latin, has increasingly in
later antiquity moved from using the middle voice to using the passive
voice and letting the passive voice do "double duty" in that the
MAI/SAI/TAI paradigms and the -QH- paradigms are fundamentally "passive"
but that relics of the older "middle" sense survive in the so-called
"middle deponent" and "passive deponent" verbs. But I no longer think that
explanation adequately explains even what has happened in Latin. Latin is
chock-full of deponents and I halfway suspect that there are more of them
in vulgar Latin than in the SERMO URBANUS of Cicero and Livy; I even
suspect that the plethora of reflexive verbs in the Romance languages
represents "the revenge of the Indo-European middle voice"--all attempts of
classical grammarians to eradicate it or drive it into exile have failed
and it has come back to haunt schoolboys (and schoolgirls) today with
self-replicating hydra-forms of reflexive verbs.
>These are not final answers, but suggestions and hypotheses worth exploring.
Thanks very much, Iver, for a very thoughtful and provocative response, and
certainly a very helpful one. Since 25K is already an interminable message,
I'll postpone my comments on APOLLUMI for another day, for fear that the
fate suffered by Euripidean prologues at the hands of Aeschylus in
Aristophanes' Frogs may come back to me: sooner or later, after five lines,
four lines, three lines, or even a half-line introductory phrase, I'll get
hit with that Aeschylean predicate, LHKUQION APWLESEN.
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