Deponent verbs are often defined as verbs that have passive forms but active meanings. But how accurate is this typical definition/generalization? It seems clear that this definition applies without problems to deponent verbs like hortor or exhortor, which do have a clear agentive/active meaning. But what about morior, orior or nascor? Can these verbs also be said to have an "active" (?!) meaning? If so, in which sense?

For example, there is a syntactic test that, in terms of usage frequency, can be claimed to distinguish deponent verbs that involve a passive meaning from the ones that encode a more active one: e.g., verbs with a clear/stronger passive meaning like morior, orior and nascor, i.a., can be easily/quite naturally found in Ablative Absolute constructions with a perfect participle, compared to more active verbs like (ex)hortari, which are perhaps not impossible but are by far more infrequent in this usage: cf. Caesare mortuo vs. (?)Caesare suos exhortato (e.g. see this link). An intuitive explanation of this difference in usage (i.e., the former is by far more frequent than the latter) comes to mind: verbs like morior or nascor involve a clear/stronger passive meaning, whereas verbs like (ex)hortor involve an active meaning. Cf. also the related meaning distinction that is involved in auxiliary selection in languages like Italian: typically, change of {state/location} verbs (patient-(derived) subject verbs, i.e. so-called "unaccusative" verbs) select essere 'be', while agentive processes select avere 'have': cf. It. Gianni è morto ('Gianni died') vs. Gianni ha esortato i suoi ('Gianni has exhorted his people').

Could you tell me if there are textbooks of Latin grammar where a more appropriate/accurate definition of deponent verbs is given? Cf. the typical one given above.

I think that TKR hit the nail on the head in his comment below.

TKR: it's an inaccurate description: "active meaning" seems to mean nothing more than "meaning which tends to be expressed with an active verb in modern European languages". To that extent the term may be pedagogically useful, but linguistically it's meaningless, and when one tries to make it meaningful as you have by translating "active" to something like "agentive", the attempt fails.

Many thanks, TKR, for your VERY useful paraphrase, which shows the confusion I alluded to above: the confusing phrase "active meaning" in the typical definition above seems to mean nothing more than "MEANING which tends to be expressed with an ACTIVE verbal FORM in modern European languages". Crucially, notice that ACTIVE here modifies (verbal) FORM, not meaning! That's why the typical definition of deponent verbs above (in particular, its reference to "active meanings") is quite confusing.

By the way, I see some of you think that the typical definition above is "pedagogically useful". Well, I don't think so... (although I admit that many years ago I learned it without questioning it).

Ethan Bierlein provides a very nice descriptive semantic typology of deponent verbs (see his post below), which can be completed with the one found in Pinzin's (2018) PhD thesis Stuck in the Middle (http://dspace.unive.it/bitstream/handle/10579/12877/956151-1197504.pdf?sequence=2 )

Basically, Pinzin tries to justify the morphology of deponent verbs by arguing that the subject is not a true/"deep" external argument, whereby, semantically speaking, it cannot be a prototypical agent, and, syntactically speaking (in formal syntax jargon), is a derived subject. In a sense, following Pinzin, one could then conclude that deponent verbs have middle/passive forms AND middle meanings. In my opinion, this definition can be said to make more sense than the traditional one above (cf. "deponents have passive forms but active meanings"), especially if, {despite appearances/despite my initial (wrong) intuition} (cf. above), the subject of verbs like (ex)hortor is not assigned a prototypical agent role. According to Pinzin (2018: 304), hortari has a meaning component that accounts for the middle/passive morphology. In his own words, the meaning of this verb is: "'x acts in such a way that x has y willing/eager (to do z)’. Such a structure would accommodate for the three objects and would provide a justification for the presence of the Middle morphology" (end quote). Very interesting, indeed! He puts it a bit more technically on page 11: "“the Middle morphology signals the syntactic absence of the external argument in the structure, may that argument be a DOER (v-doP) or an UNDERGOER (v-goP). The Nominative argument, consequently, is always merged in a low position: HOLDER of a state/location, BENEFACTIVE of a further stative event” (end quote). In my opinion, his proposal is very illuminating!

  • 3
    You're right, it's an inaccurate description: "active meaning" seems to mean nothing more than "meaning which tends to be expressed with an active verb in modern European languages". To that extent the term may be pedagogically useful, but linguistically it's meaningless, and when one tries to make it meaningful as you have by translating "active" to something like "agentive", the attempt fails.
    – TKR
    Feb 7, 2019 at 20:43
  • 2
  • 4
    I didn't know you were interested in generative syntax. If so, take a look at Embick 2000. ling.upenn.edu/~embick/latin.pdf He claims deponency in Latin is arbitrary.
    – Alex B.
    Feb 11, 2019 at 20:55
  • 1
    If one just compares two apparently very different examples of deponent verbs (e.g., agentive transitive hortari and non-agentive intransitive mori), one can easily arrive at Embick's conclusion. However, Pinzin, who, by the way, is also a generative linguist, shows that Embick's claim is probably wrong. In fact, a cursory look at Ethan Bierlein's semantic classification below is enough to realize that Embick's proposal is probably incorrect. It's true that, synchronically speaking, it is more difficult to justify the "middle" morphology but it is not impossible, as shown by Pinzin.
    – Mitomino
    Feb 11, 2019 at 22:49
  • 1
    Yes, Alex, you're right: Generative Grammar and Construction Grammar are radically different. This notwithstanding, there are many current approaches in GG that argue for a what is called a "neo-constructionist" approach to argument structure (i.e., they agree with Goldberg's CG in claiming that argument structure is not determined by the verb by rather by the construction. E.g., for a good introduction to so-called "neo-constructionism" within GG, take a look at oxfordscholarship.com/view/10.1093/acprof:oso/…
    – Mitomino
    Feb 12, 2019 at 18:20

4 Answers 4


The oft-quoted dictum "passive in form, active in meaning" is utterly nonsensical from a linguistic standpoint (and many would argue too, from a pedagogical standpoint) for describing most so-called "deponent" verbs (there do exist exceptions; these will be outlined later). The saying "middle in form, middle in meaning" captures the truth with the same succinctness; these are not "deponent" verbs: they are media tantum (middle-only) verbs.

The evolution of these media tantum verbs may at first seem perplexing: Latin has an active-passive opposition in voice; there exists no middle. However, it largely appears to be the case that these media tantum verbs arose from Proto-Indo-European middle verbs. There do exist some verbs whose origins remain murky, though. The so-called 'semi-deponents', suppletive in the perfect, are one such good example.

It is also worth noting that Late Latin saw a subsequent rise in "deponent coinages" -- which can technically be considered "true deponents" -- followed by an equally quick regularization of these true deponents and media tantum verbs towards active morphologies.

There are seven broad semantic categories into which the middle can be divided for the purpose of classifying Latin's media tantum verbs (and for affirming that these verbs are, in fact, semantically middle), those being:

  1. Direct Reflexive: an event in which the participant performs an action upon themselves. This semantic category can be expressed with the reflexive marker , but there do exist verbs in the middle under several sub-categories, typically "body action middles": grooming, bodily movement, etc.

    • ornor - to adorn (oneself)
    • perluor - to bathe (oneself)
  2. Indirect Reflexive: an event in which the participant performs an action for their own benefit. Similarly to the direct reflexive, this semantic category could be expressed with the reflexive marker sibi, but there do again exist verbs in the middle under this category.

    • liceor - to acquire (for oneself) by bidding
    • apiscor - to get, acquire (for oneself)
    • potior - to get possession of (for oneself)
  3. Naturally Reciprocal Events: reciprocal events in which participant A is performing an action on participant B, and participant B is performing the same action on participant A; i.e, events which naturally involve reciprocity.

    • osculor - to kiss
    • conflictor - to fight
    • amplector - to embrace
    • luctor - to wrestle
    • altercor - to wrangle
    • copulor - to be joined
  4. Collective: an event which is similar to a naturally reciprocal event, but the action is instead carried out by the participants as a whole; i.e, the participants are not highly distinguished from each other.

    • misceor - to mix
    • congregor - to assemble, congregate
    • colloquor - to converse, discuss
  5. Chaining: an action in which participant A acts on participant B, B on participant C, C on participant D, D on participant E, and so forth. There are very few verbs which encode this meaning, save for the obvious exception: sequor - to follow.

  6. The Cognitive Middle: a fairly broad category that can be tersely summed up as a mental event in which the subject is both the initiator and the affected; this category can be split into several further sub-categories.

    • misereor - to feel pity, pity
    • vereor - to respect, revere; fear, dread
    • meditor - to think, reflect upon
    • interpretor - to explain, expound
    • comminiscor - to devise, contrive
    • polliceor - to promise
  7. Spontaneous Process: an event in which one subject undergoes a change of state with no specified agent; the subject is the nominal participant.

    • morior - to die
    • scindor - to tear, split
    • nascor - to be born

To summarize all of the above:

  • The concept of "deponency" and the term "deponent" are not useful for explaining Latin verbs that exist without active morphologies; they are better referred to as middle-only or media tantum verbs.
  • The Latin media tantum verbs are largely derived from historically middle Proto-Indo-European verbs; semi-deponent verbs, suppletive in the passive, represent the main exception.
  • There do exist "deponent coinages" -- verbs which can truly be considered deponent -- which exist outside the class of media tantum.
  • The media tantum verbs can be divided into seven broad semantic categories, which themselves clarify why the classification of middle makes the most logical sense.

The bulk of this answer was largely paraphrased from Seumas Macdonald's excellent piece on the Latin (and Greek) middle voice: Reconceiving the middle voice for Greek and Latin students, itself largely based on Suzanne Kemmer's work The Middle Voice and Rutger J. Allen's dissertation The Middle Voice in Ancient Greek. A study in Polysemy.

While I believe that I have, for the most part, given an effective summary of the work he has done, I highly recommend that you read his original work. A PDF of it can be found here; the original blog posts from which the piece was compiled can also be read on his website The Patrologist.

An interesting aside: the accurate re-classification of so-called 'deponents' as media tantum verbs is not recent; from George Choiroboskos' 9th century commentaries on the canons of Theodosius of Alexandria:

"οὐδεὶς γὰρ λέγει ἔρχω ἢ εὔχω ἢ πέτω ἢ δέχω ἢ ὀρχῶ καὶ τὰ λοιπά, ἐπειδὴ τὸ σημαινόμενον κωλύει."

"For no one says I 'go', or I 'pray', or I 'fly', or I 'receive', or I 'depart' and the rest, since the meaning prohibits it."

  • 2
    What is the value of calling them "middle in form"? Aren't they always inflected the same way as passive verbs (or in a few forms, the same way as active verbs)? The phrase "middle in form" seems to me to imply that Latin had three distinct sets of conjugated forms for active, passive and middle, but I don't think that is true.
    – Asteroides
    Feb 11, 2019 at 6:10
  • 2
    "Middle in form" might be true etymologically, then, but is it true synchronically in Latin itself?
    – Asteroides
    Feb 11, 2019 at 6:17
  • 1
    Ok, so what about verbs like populor, cunctor, tueor, largior, etc? How would you classify them and why? Would you call populor or tueor media tantum? How's their subject "physically or mentally affected by the event expressed by the verb"? (to use Rutger Allan's definition of media tantum verbs) "cf. "Ita bifariam consules ingressi hostium fines ingenti certamine hinc Volscos, hinc Aequos populantur."
    – Alex B.
    Feb 11, 2019 at 17:12
  • 1
    @AlexB. It is indeed more interesting and useful to one approaching Latin from philology or linguistics. However, the applications of this to a classroom environment are covered in brief at the end of Seumas' work, under "For the grammar-translation context" and "Communicative Language Teaching" respectively. The application in a GT context is too expansive to summarize here (I would highly recommend a reading of the original); however application in a CI context is relatively simple: to quote "You simply introduce them without comment," students can inquire further on an individual basis. (2) Feb 11, 2019 at 18:21
  • 1
    @AlexB. No, I am not attempting to assert that all so-called 'deponents' are media tantum; there do exist Late Latin coinages which can be classified as 'true deponents' (this should be edited into my answer now). I will concede that my answer in its original state was perhaps not terribly clear about this; I suppose that's what I get for pounding this answer out at 1:00 AM. I appreciate all your inquiry towards improving clarity! (3) Feb 11, 2019 at 18:24

Summary: It seems that you and the grammarian interpret the word "active" differently, and that seems to be the source of all confusion. The defining feature of deponent verbs is the semantics of the subject, and I could well call this "active meaning". It has nothing to do with how active some action is in practice.

It seems that we have three relevant dichotomies here: deponent–non-deponent, transitive–intransitive, and active–passive. By the last one I mean "active" verbs that describe an action (like jumping or speaking) and "passive" verbs that describe a state or change thereof (like being red or being born). This has nothing to do with active and passive forms of a verb. I assume this is what you mean by active and passive (meaning) in your question; please clarify if I misunderstood.

The three dichotomies are mostly unrelated. The only restriction seems to be that a passive verb cannot be transitive; I cannot quite fathom what a transitive passive verb could mean. The other combinations are possible:

  • deponent transitive active: sequi
  • deponent intransitive active: loquor
  • deponent intransitive passive: nascor
  • non-deponent transitive active: trahere
  • non-deponent intransitive active: fluere
  • non-deponent intransitive passive: rubere

Notice that I speak of active and passive meaning, not active and passive forms. In this answer I will only treat active forms of non-deponent verbs and passive forms of deponent verbs (only forms in the "principal voice" of any given verb, if you will). I would not confuse passive forms into this discussion. Passivization of verbs with two objects might be of interest.

I see no connection between deponency and activity in either direction. There might be correlations, as the six classes above are not equal in size, but I wouldn't draw conclusions from it.

Lucanus in Bellum civile uses locuto Caesare. I think loqui is rarer in this use than mori, but I can't see why it'd be impossible. Similarly agmine secuto of Annius Florus looks like an absolute ablative of a transitive deponent verb. Absolute ablatives of certain kinds of deponent verbs might well be rare, but I cannot see why they would be impossible.

I think the point is that you have misunderstood "active meaning". Passive forms of non-deponent verbs work so that the semantic object is the syntactical subject, whereas in active forms semantics and syntax agree. For deponent verbs all forms are passive (apart form the present participle), but still the syntactical subject is the semantic subject. This relation between semantic and syntactical subjects is what "active meaning" refers to, not "activity" in the sense of making a conscious decision to act or anything like that. The bold-faced part above is what I would give as a more proper definition of a deponent verb, but I might certainly summarize it less rigorously as "deponent verbs have passive forms but active meaning".

In the definition of a deponent verb you cite, "active meaning" means that semantics and syntax agree on the subject like they do for active forms of non-deponent verbs. This can indeed be misleading or confusing. The point here is that "active" is related to active and passive forms, not "being active".

  • There seems to be a terminological confusion on what "active meaning" is (as opposed to “passive meaning”). In my view, the meaning of mori is by far less active than that of loqui, which, as noted above, probably explains why it is easier to find the verb mori in AA constructions with past participles compared to loqui: putting both verbs into the same bag (by saying that both express “active meanings”, which is incorrect in my view) does not account for this difference in usage). That's why I find the typical/simple definition of a deponent verb confusing and misleading.
    – Mitomino
    Feb 7, 2019 at 17:57
  • 2
    @Mitomino I think the point is that in "active meaning" the word "active" is supposed to refer to active forms of non-deponent verbs, not activity in an everyday sense. I agree that it can be confusing. I think you misanalyzed the word "active" (from the point of view of someone defining deponent verbs the way you cite), but I hope my answer cleared at least some of that confusion.
    – Joonas Ilmavirta
    Feb 7, 2019 at 18:02
  • 1
    @Mitomino I would say that in your example the verb docere has active meaning (in the sense you seem to understand it), but it is now used in a passive form. Verbs with two objects fall outside what I presented in this answer, but they are described in another question. Anyway, my point stands that the defining feature of deponent verbs is the semantics of the subject.
    – Joonas Ilmavirta
    Feb 7, 2019 at 18:15
  • 1
    I think the difficulty with this analysis is how to define "semantic subject". In modern linguistics "subject" and "object" are generally defined more by syntax than semantics, while semantics works with terms like "agent", "patient", etc., which don't necessarily match up with any particular syntactic roles.
    – TKR
    Feb 7, 2019 at 20:49
  • 1
    In short, so-called “unaccusativity” is what connects deponent verbs of the mori type with passive constructions. Finally, let me say that, of course, I’ve deliberately “read too much” into the notion of ‘active meaning’ in order to show its incoherence in any reasonable sense one can give to this phrase (e.g., the typical sense, e.g., the one shown in TKR’s paraphrase, or the other one I’ve also criticized in this post). Anyway, many thanks for your arguments and above all for your very interesting examples locuto Caesare and agmine secuto.
    – Mitomino
    Feb 10, 2019 at 18:01

"Active meaning" or "active in meaning" does seem to be an unfortunate expression. I agree with Joonas that it is a misinterpretation to think of this as being related to "change of state" or anything like that.

I wrote a bit about my thoughts on the "deponent" terminology in my answer to this ELU question, but basically, I think the clear thing about deponent verbs is just that they are passive in form*, and lack "corresponding" active forms. If you think of the semantic meaning of an ordinary passive verb as being derived from the meaning of the corresponding active form (e.g. you take the object of the active verb and turn it into the subject of the passive), then it's clear that deponent verbs are not "passive in meaning": their meaning can't be derived from the "corresponding active form" because that doesn't exist.

And if we think of all verbs as being either "passive" or "active" in meaning, then by elimination, deponent verbs must be active in meaning. This is very vague and doesn't require them to have any narrower a range of meanings than non-deponent active verbs (some of which are stative, and some of which are not).

*I don't think Latin has a distinction between "passive" and "middle" morphologies

By "passive in form", I am just referring to the "r" set of inflectional endings. It seems that some (e.g. the Pinzin thesis that you linked to) prefer to refer to this set of forms as "Middle" morphology. I don't think terminology is very important, but my own opinion is that, since there only seems to be a two-way contrast in this area of Latin verb morphology, it's a bit confusing to refer to one of the two sets of forms as "middle". Because the everyday meaning of the word "middle" refers to something between two other things, I had the misimpression that the use of this term implied a three-way active-middle-passive contrast, but Pinzin in fact seems to treat Latin morphology as having a two-way active-middle contrast: he says on p. 27 that "I reserve the term Passive only to those morphologies that appear only in a passive structure, in which the external agentive argument is syntactically demoted but interpretatively still present and identified with an element that is external to the derivation. Following this definition, then, Latin does not have a Passive morphology".

I understand that some people argue that "passive" is confusing because the verb forms called "passive" in certain modern languages, like English or German, don't behave the same way as the Latin form, but to me it seems better to just continue to use "passive" in reference to Latin while acknowledging the difference in meaning/usage from the English "passive".

If you prefer the term "middle", you should just be able to replace "passive" with "middle" in the preceding section.

  • We agree that "active meaning" is an unfortunate expression. We agree that "active meaning" cannot mean what this phrase seems to mean literally (e.g., active meaning does not mean agentive meaning). But then WHAT is "active meaning" supposed to mean in the typical definition of a deponent verb? My impression is that this phrase is absolutely confusing and the best paraphrase I've seen here is the one provided by TKR in a comment, where, ironically enough, ACTIVE is understood to modify (verbal) FORM, but not MEANING. See the EDIT part of my post above.
    – Mitomino
    Feb 11, 2019 at 18:08
  • As for your point ("If you think of the semantic meaning of an ordinary passive verb as being derived from the meaning of the corresponding active verb (e.g. you take the object of the active verb and turn it into the subject of the passive), then it's clear that deponent verbs are not passive in meaning" [end quote]), I think that there is indeed a parallelism between the derived/Patient subject of passives and the derived/Patient subject of unaccusative deponents of the mori type. And, yes, in both cases it seems quite natural to conclude that their meaning is passive.
    – Mitomino
    Feb 12, 2019 at 1:01
  • 1
    Putting recent literature aside, I also agree with you that in Latin there is no distinction between "passive" and "middle" morphologies. Of course, this does not prevent a construction like Porta fracta est from being interpreted in the two following ways: as a verbal passive construction ('the door was broken by someone') and as an anticausative/"middle" construction ('the door broke'). Also as an adjectival passive ('The door is broken'). Cf. a detailed discussion in latin.stackexchange.com/questions/8815/…
    – Mitomino
    Feb 12, 2019 at 1:41

Bierlien's answer is excellent, but I have something to add. In the course of puzzling out the word populabantur I noticed that the Loeb translator rendered it not as "they were devastating" (a passive word used actively) but as "they were engaged in devastating" which is a slightly different meaning and on reflection it seems that we can very often render deponent verbs the same way: the subject is engaging in doing X. So, it refers to the actor's state of being rather than to their action itself.

It is interesting to ask whether, for some words, there is any difference between the deponent and the passive voice at all. For example, populabantur has exactly the same form both as a "deponent" and as the passive voice of populabant. The Romans, as far as I can tell did not discriminate deponent verbs, and the term itself seems to have been invented by Priscian (5th century). So, the question would be does populabantur ever mean "they were devastated" (passive), or does it always mean "they were engaged in devastating"?

To answer this question I looked at every classical use of populabatur and populabantur and in every case it was used actively. So, we could potentially argue that the passive voice of populo is purely the deponent verb and there is no difference between populo and populor. Populor appears to simply be the passive voice of populo, not a separate verb. To prove otherwise, you would have to show an example of populo used passively, which I could not find.

If we look for instances in which the idea of devastation is used passively, inevitably the word vasto is used. Supposedly there is a deponent verb, vastor, but I could find no instances of its use. In every instance I could find the passive tense of vasto is used passively, not actively. Thus, it is the reverse of populo. So, I wonder how the Romans would know that populo can only be used actively and vasto cannot be used as a deponent (if that is so)? There must be some subtle aspect to understanding the voice of the verb that would tell them this, because to simply memorize this for every possible verb seems too difficult.

  • The interesting translation you give of populabantur 'they were engaged in devastating' has reminded me of Francesco Pinzin's (2018) definition/classification of this denominal verb (his PhD thesis is downloadable at dspace.unive.it/bitstream/handle/10579/12877/956151-1197504.pdf ). See page 121: The nominative argument is in control of the event (it is a 'doer') but, in order to justify the middle morphology, Pinzin adds the following: this nominative argument is also "the benefactive of a stative event in which another argument (marked by Accusative case) is involved".
    – Mitomino
    Dec 3, 2023 at 18:10
  • By the way, when dealing with Latin deponent verbs, one thing I've always found intriguing is why the most productive set of them consists of denominal and deadjectival verbs (see page 119 and passim).
    – Mitomino
    Dec 3, 2023 at 18:13
  • 2
    Surely the Loeb translation is merely rendering the imperfective aspect of the tense, and it has nothing to do with the verb's being deponent? There are different ways of rendering this aspect for a past imperfect verb form: 'they were xing,' 'they were in the process of xing,' 'they xed (regularly/habitually),' and even 'they began xing' and 'they tried xing.' Your translation is just a variant of the 'they were in the process of xing' pattern. After all, I could just as easily – and, I would maintain, validly – translate the nondeponent populabant as 'They were engaged in devastating.'
    – cnread
    Dec 3, 2023 at 20:15
  • @cnread You're quite right but don't get me wrong. I've just noted that this translation ("engaged in") has reminded me of Pinzin's (2018: 121) classification of the verb populor, in particular his non-trivial/non-obvious point: "the benefactive of a stative event in which another argument (marked by Accusative case) is involved" (italics mine).
    – Mitomino
    Dec 3, 2023 at 20:25
  • 2
    "To prove otherwise, you would have to show an example of populo used passively, which I could not find." -- Lewis & Short list several instances (no finite forms, mostly part. perf., which is typical, e.g. commentus = "invented"). I do think the whole example "agros populabantur" is a good demonstration of the truth in the formula "active meaning, passive form." Dec 3, 2023 at 23:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.