Light verbs and suicide

If you go to PubMed/MEDLINE, the US National Library of Medicine’s giant repository of (and search engine for) biomedical publications, and look around for papers on language and suicide, you won’t find that much on what you’re probably expecting: research on the language of suicidal people.  What you will find is papers on how we talk about suicide.

The major issue has to do with the ways that we refer to the act of suicide.  In English, your basic options are:

  • to commit suicide
  • to kill oneself
  • to take one’s own life
  • to do oneself in
  • to die by one’s own hand

The problem is that first one: to commit suicide.  People who work in the field of suicide in any capacity–prevention, treatment, research, whatever–aren’t very fond of it.  The reason: it stigmatizes the act.  In English, things that you commit are bad.

Now, you’re thinking: commit isn’t always bad, right?  You can commit to doing something, commit to someone, commit something to memory.  No question!  But, we’re seeing two very different linguistic phenomena here.  The bad commit has a very specific kind of structure: it’s what we call a light verb.  

Light verbs are a special kind of verb.  They don’t have very much meaning by themselves.  Rather, they occur with some other verb, and it’s that verb that gives the expression its meaning.  For example: in English, the verb to take can be a light verb.  (It isn’t always a light verb–but, there are many times when it is.)  Here are some English-language expressions in which to take is a light verb:

  • to take a bath
  • to take a beating
  • to take a break

What does to take mean in to take a bath, to take a beating, and to take a break?  I suggest to you that it doesn’t mean very much at all.  Rather, it’s bath, beating, and break that contain the meanings of those expressions.  (I’ve put a technical definition at the end of the page, if you are into that kinda thing.)

Light verb constructions are not a rare phenomenon.  Here are a bunch more expressions in English that are what we call light verb constructions (that means the light verb plus whatever it is that it combines with–in English, typically a noun) in which the light verb is take:

  • to take a breather
  • to take a bus/taxi/shuttle/plane/train
  • to take a dump
  • to take a gander (at)
  • to take a minute
  • to take a pee
  • to take a piss
  • to take a shit
  • to take a vacation
  • to take a walk
  • to take pity (on)

…and, it’s not like take is the only light verb in English.  In fact, we have several.  Some examples:

  • to make a decision, to make an offer, to make haste, to make peepee
  • to give a shit, to give (someone) a hand, to give a damn, to give a fuck, to give birth (plus some obscene ones that I’m leaving out)
  • to get dressed, to get ready, to get angry/mad, to get nasty, to get drunk, to get high, to get sober
  • to do battle, to do business, to do your business (yes, those are different)
  • to have a ball, to have a blast, to have fun, to have a good time, to have a headache, to have mercy, to have sex, to take a piss
  • to take action, to take a seat, to take one’s time, to take note, to take notes (yes, those are different), to take a look (at)

With that data in hand, we can see the difference between the commit of to commit suicide and the non-bad senses of commit in to commit to memory, to commit to a person, to commit to a deadline…  none of those have that verb + noun structure.  They’re all commit + to something.

Some more (or less) useful light verb constructions in French: faire la vaisselle: to do the dishes faire la lessive: to do the laundry faire [+université]: to go to a university (J’ai fait William and Mary, I went to William and Mary) faire du [+musical instrument]: to play an instrument (as in to do so habitually) faire du diabète: to have diabetes





One of the interesting things about light verb constructions is that they don’t show a basic characteristic that we expect to see in language: compositionality.  To paraphrase from a previous post:

Compositionality is the process of meaning being produced by something that you could think of as similar to addition (technically, it’s a more general “function,” but “addition” will work for our positions–linguists, no hate mail, please).  Take a situation where my dog stole some butter.  The semantics are: there’s a dog, it’s my dog, there’s some butter, and the butter was taken, by the dog, without permission.  (You can’t believe how horrible the poo that I had to pick up over the course of the next 24 hours was.)  My dog’s name is Khani, so I might say something like this: Khani stole some butter.  The idea behind compositionality is that the meaning of Khani stole the butter is the adding together of the meanings of Khani, steal, butter, and the meaning of being in the subject position versus the object position of an active, transitive sentence.

So: we have this basic expectation that meaning in language will be compositional, and as linguists, as computer science people who work with human language, and as philosophers, we have a hell of a lot riding on that expectation.

In that context, the cool thing about light verb constructions is this: they’re not compositional.  There is pretty much no way to get any systematic interpretation of the combinations of light verbs and their nouns.  Pause and ponder:

  1. To make peepee and to take a piss: they mean the same thing (the difference is that one is child language and the other is too impolite to say in front of your grandmother).  Peepee and piss mean the same thing–again, one is child language, and the other one is too impolite to say in front of your grandmother.  Your assignment: tell me what make and take contribute to the meaning of those expressions–that is, explain to me what their contribution is to the composition of the verb and the noun.  My point: peepee and piss mean the same things, and to make peepee and to take a piss mean the same things–how do you explain that, if make and take each contribute something to the meaning of those expressions?
  2. Consider to take a bath and to take a bus.  One of those is what you might think of as event–an act of bathing.  The other is a big, smelly thing that takes you to work.  (See how I slipped another take in there?  Different take–nobody said that linguistics was going to be easy.)  In take a bath and take a bus, your relationship with the two things is pretty different.  In the first case, you’re participating in an event, while in the second case, you’re making use of a mode of transportation.  Your assignment: tell me how that difference comes from the verb to take.

The answers:

  1. Trick question: as far as I know, make and take don’t contribute anything to the meanings of those expressions.  The meanings of the expressions are not compositional.
  2. Trick question: as far as I know, take doesn’t contribute anything to the meanings of those expressions.  The meanings of the expressions are not compositional.

So, back to to commit suicide.  As you might have noticed, the relationships between light verbs and their nouns are things that a child learning their native language just has to remember.  There’s nothing that the kid learns about their language that would let them infer or guess that they’re making peepee now, but they’ll be taking a piss when they grow up: they have to remember it when they’re exposed to it.  (I don’t mean to suggest here that children learn language by remembering stuff to which they’ve been exposed–we’ve talked about how very little of language-learning for children works that way.)

So, the verb + noun combinations in light verb constructions are pretty random.  The thing about commit is this: it’s a light verb, too, in constructions like to commit suicide.  Its noun is of a very specific kind, though: its verb is something bad.  Compare that with the light verb to have.  You can have a heart attack, you can have a migraine, or you can have a good time, or have sex.  No particular semantic consistency there–could be bad (heart attack, migraine), or it could be good (a good time, sex).  Here’s a list of the words that are statistically most closely associated with the verb to commit in the enTenTen corpus (a collection of 19.7 billion words of written English, available on the Search Engine web site; git is a computer science thing.  See this post.)

Top objects and subjects of the verb “commit” in 19.7 billion words of English. From the enTenTen corpus on the Sketch Engine web site.

What kinds of things get commited?  Crime, sin, murder, fraud, atrocity.  Who commits things?  Offender, defendant, criminal.  Not good.

So, what are we to make of to commit suicide?  Many people who work in the field (go do your own search on PubMed/MEDLINE if you’re interested, or just see here and here for examples) are of the opinion that use of the expression to commit suicide has the effect of stigmatizing the person who killed themself.  Is that a bad thing?  They think it is.  I think it is, too.  Now, does the person who killed themself care how you talk about them?  Certainly not–they’re dead.  But, that person’s mother, husband, son, daughter, cousin, aunt, uncle, best friend…  They do, and they’re not dead.  So, many people who work with suicide in some capacity would like to see that expression go away.  Here’s a very eloquent expression of the idea, from Doris Sommer-Rotenberg:

The expression “to commit suicide” is morally imprecise. Its connotation of illegality and dishonour intensifies the stigma attached to the one who has died as well as to those who have been traumatized by this loss. It does nothing to convey the fact that suicide is the tragic outcome of severe depressive illness and thus, like any other affliction of the body or mind, has in itself no moral weight.   —Doris Sommer-Rotenberg, Suicide and language

Who cares?  Sommer-Rotenberg again:

The rejection of the term “commit suicide” will help to replace silence and shame with discussion, interaction, insight and, ultimately, successful preventive research.

Now, I know what you’re thinking: you’re thinking that the language that we use doesn’t affect the way that we think.  You know what?  I agree with you.  However: even if how we talk about suicide doesn’t change the way that we think about it, I suggest to you–as a fellow student said to me in the basement of Oxley Hall one evening in our graduate student days: the fact that the language that we use doesn’t change our reality doesn’t change the fact that you can make someone feel bad with the language that you use, and you wouldn’t want to do that, right?  Well, of course not.  (I love my job, but my fellow ex-student’s job is definitely cooler–she’s the only speech therapist in New York City whose practice is exclusively concerned with transgendered people.)

There’s also controversy/discussion around the ways that we talk about what happens when people try to kill themselves, but don’t succeed–attempted suicide, unsuccessful attempt, failed attempt, failed suicide, and failed completion, versus completed suicide–the idea is that these expressions model suicide as a desirable act.  (If you fail to have a good time, fail to get into medical school, fail to convince someone to marry you, that’s a bad thing.) See here for a fuller discussion.

So, yeah: there’s probably more stuff in the National Library of Medicine’s repository on how we talk about suicide than there is on how people who are suicidal talk.  It would be great to change that, because if we knew more about how people who are suicidal use language, then we might be able to do a better job of preventing it.  Now, that means that we need language data from people who are or have been suicidal, right?  But, we also need data from people who aren’t suicidal–if you want to understand something, you usually need to compare it to something else, and in this case, that means comparing the language of suicidal people to the language of people who aren’t suicidal.

It happens that there’s an enormous amount of real, live language out there in the world on social media platforms.  It would be great for suicide researchers to have it, but there are ethical issues involved–just because someone puts their life out there on the web doesn’t give you the right to just grab it and do stuff with it.  However: you can donate your social media data to, a group that collects language from all kinds of people for social media research.  You can sign up with them here.  As the character Père LeFève says in Anne Marsella’s wonderful short story The Mission San Martin:

Best wishes to all of you who are still alive.  And if you’re yet alive, please give.

–Anne Marsella, The lost and found and other stories

No English notes as such today.  Instead, here’s some extra stuff for those of you who like to dive deeply into the linguistics of things.

One way of defining light verb construction:

Definition of light verb construction (LVC) from Light verb constructions in Romance: A syntactic analysis, by Josep Alba-Salas.

On the mechanics of how the meaning gets out of the noun and into the verb, so to speak:

Contrary to “prototypical” verbal constructions where the verb is the syntactic and semantic head of the sentence and its syntactic dependents are also its semantic arguments, in LVCs, one of the syntactic dependents of the verb, generally its direct object, functions as the semantic head, projecting its own argument structure, while the verb, which is semantically “light”, bears only inflection and projects no argument structure.

− Given the fact that the verb has no semantic contribution or rather its semantic contribution is quite weak, it cannot be selected lexically, that is on the basis of its semantic contribution. The combination of a particular predicative noun (PN) with a particular light verb (LV) is thus a matter of idiosyncrasy: The noun and the verb form a collocation that must be stored in the lexicon.

Pollet SAMVELIAN, Laurence DANLOS, and Benoît SAGOT, On the predictability of light verbs

Why would you need to posit the existence of such a thing?  From Samvelian et al.:

  1. l’agression de Luc contre Marie (the attack of Luc against Mary)
  2. Luc a agressé Marie (Luc attacked Mary)
  3. Luc a commis une agression contre Marie (Luc committed an attack against Mary)
  4. l’agression que Luc a commise contre Marie (the attack Luc committed on Mary)

Note: you can also se commettre avec quelqu’un.

9 thoughts on “Light verbs and suicide”

  1. The language that we use doesn’t affect the way that we think ? On this point I’m not sure . Some thinkers seem to suggest that the reality that exists in our consciousness is largely depending on the description of the world that we do, which is largely submitted to the language that was imposed to us . And by reality they don’t only mean inner states but the whole image of the universe around us . It is a very deep subject, and it belongs to Philosophy in its noblest sense I think .

    Just an exotic situation : in the Cherokee language there is the “immediate tense” which translates to “I just went” and “I am about to go”. This same tense also acts as the command form for “You” and “We”.
    This means this Cherokee tense works for past as well as future . The immediacy is the only notion that matters . I never heard of a tense like this, this implies a vision of our relationship with the universe highly different from ours .
    The imperative is very suggestive too . Nowhere did I hear of a conjugation that doesn’t make a difference between “You” and “Us” (except English of course but I only consider serious languages).
    The same tense for before and after, the same form to include and to exclude “myself”, this is incredibly alien . For me at least, and maybe for most Earthlings .

    Liked by 2 people

    1. Well put. I think about this kind of thing a lot, what with spending many of my waking hours trying to learn to speak the language of Molière (or at least the language of Queneau, which might actually be harder!).

      As linguists go, I lean pretty heavily towards the empirical. Where that comes up pretty heavily in questions of linguistic determinism (the idea that language determines thought, world view, however one wants to put it) is what happens when you try to demonstrate such a thing experimentally. To wit: I can demonstrate pretty conclusively that American English speakers contrast with speakers of many other languages with respect to where they draw the lines between the sounds p/b, t/d, and k/g. (From the point of view of how the sounds are formed, those are identical.) The difference between those pairs of sounds is the timing between when the consonant is formed and when the subsequent vowel starts. We’re talking about a matter of milliseconds. Take a bunch of native speakers of French and a bunch of native speakers of American English and play them those sounds with different time differences, take the average of each group, and it will be different for the two groups. So, yes, you could say that in a very real way, the two groups perceive the world differently, and hell–I can quantify it. The question is: other than phoneticians, who cares? We’re always looking for profound effects of things like different ways of expressing time, or having a gender system versus not having one, but nobody can demonstrate them in a way that’s particularly interesting. Given that they’re supposed to be pretty profound, one would expect them to be pretty obvious. And yet: all that I can show you is these millisecond-level differences in things like speech sound perception.

      Hence the skepticism of people like me. But, as I said in the post: that doesn’t mean that you shouldn’t avoid saying something if you know damn well that it’s going to hurt someone’s feelings.

      (I love these conversations, by the way–thanks!)

      Liked by 1 person

      1. I didn’t mean differences in timing and sounding, but in concepts, that are translated (and infered) by our wording . If a kid is brought up in a tribe where they believe in spirits in nature he will perceive these spirits where we don’t . One can say it doesn’t change the actuality or not of these spirits, but – and this is Philosophy, and even more, metaphysics- what does exist out of consciousness ? And most of the time aren’t the objects of our conscience more powerful over our lives than anything else ?
        But about actuality, I remember having been healed by a Hopi 60 y/o medicine man who, after the healing told me that I had been caught in a (?) wind . It rang a bell in me but not clearly at all, and his younger brother said “You know, this kind of special nasty winds that happen when …” . It was obvious for them, for Indians, this reality was well known, experimented and handled . They thought I was able to have access to this reality (otherwise I wouldn’t have had the privilege to be healed by this highly venerated medicine man some white people had to spend a year just to see) . These people perceive, deal with, draw theories from, other realities . When their babies start speaking they learn another way of describing, interpreting, dealing with the universe . And from what I lived, the healing was stunningly efficient and quick, so is their reality less real than ours ? Reality is basely force fed on us by our parents and co, and their weapon is their damned language . ” Oh, look at the tree !” when a baby only sees vibrating lights with no meaning . This world of mine makes me feel like suing my parents and their society !


  2. A fascinating walk through light verbs (did I know such things existed … of course not – I just use them naturally in English and make clumsy stabs in the dark at getting them accidentally right in French). This is what I love about your work …. I learn. And then I learn more. I have one little comment on the expression ‘to commit suicide’ (and like all those referenced, I hate the negative tags applied to suicide) – it was illegal in England for many years. It was considered, I think across the Catholic world and even when the Church of England was founded by Henry VIII the prejudice persisted, that only God could decide when you died therefore if you ‘took your own life’ you were a criminal and you could not be buried within a concentrated graveyard. The suicide graves, like the graves of illegitimate children, even babies, were outside the walls. At least we have moved away from that notion but it would be nice to think that we could invent some different terminology that doesn’t infer that it is still a crime as well as the most desperately tragic of ends.

    Liked by 2 people

  3. Make and take are semi-phatic. You’re analysing denotation as sole contributor to function but connotation is equally important; make has a socially constructed meaning that includes a sense of immaturity.

    If you’re right, and there’s no compositionality to make and take in “make peepee”/”take a piss”, why don’t *”make piss” and *”take peepee” work?

    Liked by 1 person

    1. (sorry, weird iPad bug) “not much” is not “none.”

      If you’ll grant me that ALL meaning is socially constructed: what does “semi-phatic” mean? I’ve often heard “phatic” used to mean things that are meant to communicate something about social relations, versus their content (“good morning!” is my classic example), but I haven’t heard of “semi-phatic.”


  4. I did some searches on Sketch Engine to see if you’re right that “make piss” and “take peepee” don’t work, and as far as I can tell: you’re right!

    Regarding no compositionality: I wouldn’t go that far, personally. I didn’t in the post, and Samvelian et al. don´t, either. “Not much” is not


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Curative Power of Medical Data

JCDL 2020 Workshop on Biomedical Natural Language Processing


Criminal Curiosities


Biomedical natural language processing

Mostly Mammoths

but other things that fascinate me, too


Adventures in natural history collections

Our French Oasis


ACL 2017

PC Chairs Blog

Abby Mullen

A site about history and life

EFL Notes

Random commentary on teaching English as a foreign language

Natural Language Processing

Université Paris-Centrale, Spring 2017

Speak Out in Spanish!

living and loving language




Exploring and venting about quantitative issues

%d bloggers like this: