Can causative and anticausative co-occur in Turkish verbal morphology?

Question

Turkish makes use of two valency markers: (i) the causative marker with 'tur' which increases valency in (1) below, and (ii) the anticausative marker 'il' which decreases valency as in (2) below.

(1)

(ben) Hasan -8     kitab -1    oku  -t     -tu   -m
I     Hasan -lld:. book  -kc.  read -Caus. -Past -l.sg.
"I made Hasan read the book"

(from Kornfilt, 1997, p.331)

(2)

Kapį aç-il-di. <br/>
Door open-ANTIC-PAST <br/>
‘The door opened.’ <br/>

(from Haspelmath, 1987, p.2)

Combining these two markers renders ungrammaticality in some languages, but I'm not sure whether ungrammaticality follows in Turkish, i.e. when we combine both of these markers. I'm not a native speaker, so I cannot tell. What I want with this is to find an authentic sentence whereby 'tur' is combined with 'il' in the same stem and see whether they render the sentence ungrammatical or not.

*I apologize for the previous way in which this question has been asked.

score 4 · Answer 1 · answered Jul 27 '19 at 10:16

4

Definitely, it's especially common for certain verbs.

Bir ev tahliye ettirildi

Translation:

A house was evacuated

In Istanbul, following the past earthquake many houses cracked (sic.). Among these, opposite of the Süreyyapaşa factory in Balat, was a new four-story stone house whose cracks were deemed so dangerous that yesterday security officers (sic.) escorted the renters outside while their belongings were transported to another location.

(Cumhuriyet Gazetesi. 4 May 1935, p. 2.)

Here tahliye is a word which means an evacuation. (I'll spare you the ultimate derivation to keep things as simple as possible)
To make it into a verb we use the word etmek. "tahliye etmek" means "to evacuate"
Let's make it causative, "tahliye ettirmek" would be to order an evacuation.
Finally what would be the fate of one who has had an evacuation ordered for them? Pop in the anti-causative -il and you get "tahliye ettirilmek"
Because this is from a newspaper and we're reporting it, you can conjugate it in the third person plural* and past tense yielding: "tahliye ettirildi"
- (* The third person plural is equivalent to the singular in a lot of cases, such as when the subjects are humans)

And that's the gist of it.

Also note that there are a few other forms of the causative as well so you might also come across pişirilmek (to get cooked) or kapatıldı (to be/get shut down).

So for instance:

Akşam için muhteşem bir yemek pişirilmiş.

A wonderful meal has been cooked for the evening.

Or:

Eski ofisimiz kapatıldı ama yenisi şu adreste açıldı.

Our old office was closed but the new one opened at this address here.

answered Jul 27 '19 at 10:16

madprogramer

735
1
5
20

1

Thank you. You're right about the newspaper language, reporters tend to hide AGENCY by using passive voice, anticausative voice, etc. I asked this question because I have noticed that morphological causative (MC) and morphological anticausative (MA) do not co-occur in Berber, while the MC and passive do occur. It's well known that passives have hidden agency (by-phrases), but morphological agentless in Berber. Passive agency can be accessed by some tests. But the AM agency cannot, meaning that it doesn't exist, are you sue that this is a MC and MA co-occurence not a MC and passive? – Tsutsu Jul 28 '19 at 10:10
"But the AM agency cannot, meaning that it doesn't exist, are you sue that this is a MC and MA co-occurence not a MC and passive?" @TsutsuT. I apologize for the late response it took me several re-reads to understand what exactly you were asking. Well I have good news and bad news. The good news is that you're kind of on the right track because the Turkish anti-causative is often called a passive (especially if you're working with Turkish materials). As for the bad one, the preference of passive over anti-causative is more so due to philological tradition. – madprogramer Aug 11 '19 at 05:40
contd. If you're asking whether or not it meets the definition of "an intransitive verb that shows an event affecting its subject, while giving no semantic or syntactic indication of the cause of the event." (Wikipedia) for MA, then yes it very much certainly does. As for by-phrases while there are some equivalents (using postpositions, and I use this word lightly, such as tarafından or vasıtasıyla or the -ca/-ce suffix) there is still tbc. – madprogramer Aug 11 '19 at 07:37
Thank you again @madprogramer. I appreciate your patience and consideration. So am I safe to conclude that Turkish anticausative and morphological causative do not co-occur? – Tsutsu Sep 02 '19 at 09:41
@TsutsuT. I don't believe I can answer your question without rewording it. Instead of trying to answer if what we see here is MC+MA or MC+Passive, I'll answer if one can produce an example such that the verb will have an agent undetectable by by-phrases etc. Which is a yes, for instance "Çiçekler açıldı" would mean "The flowers have blossomed" and yet it isn't possible to assign an agent in this case with by-phrase or similar. To attribute the process to God or nature etc. will require the agents to be described adverbially, dare I say causatively or instrumentally. – madprogramer Sep 03 '19 at 13:49
Allah'ın izniyle çiçekler açıldı or Baharın gelmesiyle çiçekler açıldı means the flowers blossomed by God's will or by the arrival of spring, but if God or nature were to be agents then one would have to remove the suffix and produce sentences such as "Allah emretti ve çiçekleri açtı" or "Baharın gelmesi çiçekleri açtı". The issue here is that what makes "açıldı" differ from "açtı" semantically is that it refers to an internal process. And yet one cannot assign the subject to be the hidden agent (although as a matter of fact this usage is lumped into another class of verbs – madprogramer Sep 03 '19 at 14:01
often described as "reflexives"). The issue here is a deep one, as both national and worldwide academia are working with controversial definitions but to summarize: this is an internal action, where we are unable to present a hidden agent and is morphologically identical to what some dub as the passive and others as the anticausitive. I would suggest you read
IS NON-ACTIVE MORPHOLOGY A RELIABLE INDICATOR OF EXTERNAL CAUSATION? EVIDENCE FROM TURKISH to get a better idea – madprogramer Sep 03 '19 at 14:07
for why a case is made that the Turkish "passive" is at least in some contexts able to behave as an anticausative. So I don't about Berber enough to translate your analogy; but instead of having separate forms for MA and Passive, Turkish has a set of suffixes which can give the effects of MA, Pacification and even Reflexivity and it is somewhat up to interpretation to decide how distinct these are on a semantic level. As arguably interpreting all of these to be MA solves the problem of missing hidden agency and interpreting all of these to be Passive forms makes things simpler 90% of the time – madprogramer Sep 03 '19 at 14:11
In conclusion many scholars by far have chosen to analyze these as MC and passive, historically. But as the concept of anti-causativity has become more and more well known among academia it is providing a "transitory phase" to see what's going on in that 10% or so of cases where the verb form can barely can be called a passive. If you ask me it's probably neither MA or Passive by either definition and to my knowledge it has been compared with what is described as a "middle voice" or "middle-passive voice" which appears in Greek and Albanian. The problem here is sadly an ontological one – madprogramer Sep 03 '19 at 14:16
and for research purposes I think you're safe to take it as either MA or Passive. For pragmatic purposes you can go with whatever floats your boat. Your original question of "causative and anticausative" co-occurence in Turkish verbal morphology is a big yes. Whether you decide to call it a passive shouldn't change that :) So I can only hope, I mean it when I say only hope, that this was a sufficient answer. If you don't have any more qualms would you now consider accepting it? – madprogramer Sep 03 '19 at 14:19
Thank you again for your time :). Let me clarify some points. First, apart from ACTIVE, there are many voice categories that reduce agency, e.g., anticausative, passive, middle, reflexive, reciprocal, etc. etc. I'm aware that agency is reduced, but it doesn't mean it's a passive here, at least this is what Haspelmath (1987) belives. That there is a co-occurence of morphological causative and passive is clear and well-known. But ungrammaticality follows when an anticausative co-occur with a morphological causative, in Berber I asked myself why? and I could stipulate that the 's' marker of MA – Tsutsu Sep 04 '19 at 11:56
in Berber does one important function: empowering light v. Because in the case of anticausative, the vP is defective (i'm using Phase Theory as a framewrok here, Chomsky (2000,2001,2008)). When 's' is inserted, it empowers v which becomes a v* (v star) meaning it needs an EXTERNAL ARGUMENT (i.e. a subject). And we made it clear that anticausative does not have agency whatsoever unlike passive. That's why the hidden agency in the passive can be accessed either by 'by-phrase' (not in Berber, but in English), and since Berber has an agentless passive, the agent can also be accessed by the 's' – Tsutsu Sep 04 '19 at 11:59
marker of the MA. Therefore, it's impossible for a voice to be called at the same time PASSIVE and ANTICAUSATIVE. This means that the data you provided can be either a co-occurence of MC & Anti or MC & Passive. I would like you to do a last step for me, please. I'll provide data from Berber, and base upon it to show either grammaticality of ungrammaticality. – Tsutsu Sep 04 '19 at 12:01
Data: from Berber (spoken in Northern Morocco): yčaz yijjar 'the field was ploughed': this is ANTICAUSATIVE, when combined with morphological causative it yields: *y-s-čaz yijjar 'he made the fild ploughed' which is UNGRAMMATICAL. Now, the passive tw-azrn idamn 'the blood was shed' can be combined with morphological causative 's' yielding tw-a-s-izrn idamn 'the blood was made to be shed', which is totally GRAMMATICAL. Can you provide the same data to compare them? Thank you so much for your help. – Tsutsu Sep 04 '19 at 12:01
by MA I was referring to MC, I apologize, MC=Morphological Causative with 's'. – Tsutsu Sep 04 '19 at 12:08
Let us continue this discussion in chat. – madprogramer Sep 04 '19 at 12:45
@TsutsuT. unfortunately the room was frozen before I could write my answer and somehow it deleted my messages after I tried adding you to a new room and you didn't join on time or something XD In summary it depends on the verb. "Eğ-dir-il-mek" to refer to someone who has been forced to bow is rather common but "Aç-tır-ıl-mak" to refer to someone getting something to open feels redundant as you can just "aç" it in the first place. Both are definitely grammatical, but the usage will vary. – madprogramer Sep 20 '19 at 06:55
Sometimes I imagine "yaz-dır-ıl-t-tır-ıl-dı" and other hypothetical chaining examples and I suspect people would be able to make sense out of this no matter how unnatural it might sound. The preferred way to chain agency is often to break with another verb and add agency information into there. "Yaz-dır-ıl-t-mış Ol-dur-ul-du" which again while sounding unnatural is again acceptable, in theory. – madprogramer Sep 20 '19 at 06:58
1

I think I'll conclude this with
"O Eğ-di" (He bent something? What did he bend I don't know) [3rd P. Sing.] [Bent][-di Past Tense Marker][no conjugation] ||| "O Eğ-il-di" (He was bent, that is to say he bowed. Somehow. I can't assign a proper agent.This is a nice example for supporting the anti-causative paradigm over the passive paradigm) [3rd P. Sing.] [Bent][ANTIC][-di Past Tense Marker][no conjugation] ||| "O Eğ-dir-il-di" (He was made to bow. Who made him bow? I could just assign a "by-phrase" if I must ) [3rd P. Sing.] [Bent][CAUS][ANTIC][-di Past Tense Marker][no conjugation]
– madprogramer Sep 20 '19 at 07:00
Was that any better? – madprogramer Sep 20 '19 at 07:03
thank you so much. Now, you've proved that the causative -dir- and anticausative -il- can co-occur in Turkish. Let's not argue further about this fact. But let me ask one thing, is this occurence motivated by the general outlook considering that there's some kind of agency with -il- or is it the case that the causative -dir- can co-occur with a passive affix as well? I mean, is -dir- restricted in its occurence with anticausative -il-or it can freely co-occur with the passive affix as well? – Tsutsu Sep 21 '19 at 11:51
I'm not sure I follow? As far as Traditional Grammar is concerned my derivation is wrong and the 3rd form is a passive rather than an anti-causative. So [3rd P. Sing.] [Bent][CAUS][PASSIVE AFFIX][-di Past Tense Marker][no conjugation]. But again, that's the whole issue since grammarians can't seem to agree on the nature of -il- as a Pacifying or Anti-Causitivizing Affix. Ok I get it, in many languages a distinction is made, but in the case of Turkish I'm pretty sure most people who study the Grammar don't even know what an anti-causative is. So could you try clarifying that question a bit? – madprogramer Sep 25 '19 at 07:09
Apologies, I wasn't clear on my question. What I understood from your comments is that the -il- can co-occur with -dir-. Let's us say that -il- is anticausative. And let's say that -dir- is causative. Now, the passive in Turkish, I don't how it's marked, let's say it's marked by ABC affix, can -dir- and ABC affix co-occur as -il- and -dir-? – Tsutsu Sep 26 '19 at 10:03
1

Ok, and I'm telling you the ABC suffix is -dir. Based on how 99% of the agglutination in Turkish occurs by post-appending suffixes, the "traditional" logic to explain this would be that -dir (deliberately not -dir-) has "pacified" the verb eğ-dir-mek. There's not much room for further discussion here. I think what you might actually be interested in could be the nuance in the preference of -n (rare that a verb should be able to take both but eğmek will suffice, eği-n-mek to "tend" or display a "tend"ency) or -r (eğirmek, to spin/loom, but this is often taken to be a separate root) over -il. – madprogramer Sep 27 '19 at 12:19
1

To be honest I don't think there's an agreement on this. The "rules" are that it's a "-n" for verbs ending in vowels and "-il" for those ending in vowels with the "-r"s being irregulars. You will however sometimes see examples like this were they don't quite fit this as well as a lot of "-nil-" where they follow one another but this is often shrugged off as misspeech. If you are still insistent on searching for such an ABC suffix than I'd recommend you start by reading up on these alternative forms your self, but again "no agreement" so don't expect to find your answer here either. – madprogramer Sep 27 '19 at 12:35
Anyhow I digress, if you want the answer to this either ask:
1. Do all languages have/Does Turkish have a distinction between the passive and anticausative
2. What are the differences between the Turkish passive voice suffixes. Anyway I'm hope you've enjoyed this 2 month long conversation XD
– madprogramer Sep 27 '19 at 12:46
Thank you so much for your patience. It's been a long way as you said, but I benefited from your insights. I'll try to read more references about this issue. I wish you the best of luck in your studies :) – Tsutsu Sep 28 '19 at 08:39

Can causative and anticausative co-occur in Turkish verbal morphology?

1 Answers1