Expressive verbs in Tuareg

Some years ago I was following a course by Maarten Kossmann on Tuareg (Tamasheq, Tamajeq, Tamahaq). It was thoroughly enjoyable. After the first lecture we were all alotted a letter of the great Dictionnaire Touareg — Français1 (a consonant, obviously), and for the remainder of the course these 15 to 40 dictionary pages would form the basis for a number of excercises in the weeks to follow. It was an interesting method; for Kossmann this must have felt like letting loose a bunch of sheep just to see what patterns their random grazing would produce.

Anyway, I was given the letter F and I remember being intrigued, during those hours of grazing pp. 169-189 of the Dictionnaire, by those classes of verbs that were characterized by reduplication. There are various types of reduplication here: classes 7, 8 and 15 are characterized by full, classes 11 and 16 by initial, and class 9 by final reduplication.2 The interesting thing is that these verbs also seem to cluster together semantically in various ways.

Examples in F

Let’s look at some data. Of the 168 verb roots starting in F in the dictionary, there are five class 7 verbs, eight class 8 verbs, 18 class 9 verbs, and two class 11 verbs (no verbs from class 15 and 16 in this sample). That’s 5%, though that percentage is pretty meaningless with such a small sample. Here’s a selection of expressive verbs:

class verb definition
7 fələɣfələɣ lamper à grand bruit [gulp with much noise]
  fənəgfənəg se dandiner (marcher comme un canard) [waddle like a duck]
  f̣ərəsf̣ərəs couper en plusieurs morceaux, hacher [cut into several pieces, mince]
  fərətfərət germer çà et là [sprout up here and there]
  fərəzfərəz asperger / éclabousser [sprinkle, splash]
8 făẓfăẓ broyer, écraser, casser en tout petits morceaux [crush, pound, pulverize]
  făkfăk ê. très léger par rapport à son volume [light as a feather]
  nəfədfəd trembler tellement3 [tremble heavily]
  nəfəyfəy délirer, déraisonner (malade fiévreux)4 [rave, be talking nonsense (feverish illness)]
9 f̣əkənkən marcher/grimper/courir [walk/climb/run]
  fələzləz 1. verdoyer [become green] 2. fleurir [bloom, blossom]
  fələɣləɣ rendre un son faible et cristallin en coulant [render a soft and cristalline sound]
  fərəɣrəɣ 1. faire un bruit fracassant [make a thundering noise] 2. cliqueter [be clicking]
  f̣ərəkrək 1. bouger [stir, move] 2. produire un petit bruit de craquement [produce a small cracking sound]
  fətəɣtəɣ couler à grand débit/à flots [flow in great amount/with waves]
11 f̣əf̣f̣ərət frotter [rub]
  fəẓəẓẓəkət ê. extrêmement chaud (jour/soleil etc.) [be very hot (of a day, of the sun, etc.)]

Beautiful words, don’t you think? My personal favourite is fələɣləɣ ‘render a soft and cristalline sound’, though fərətfərət ‘be germinating here and there’ is a close second.

Let’s take a closer look at these verbs. Prasse calls them ‘expressive’, but doesn’t go into their semantics at all.5 Kossmann, in an unpublished sketch of Tuareg, devotes some paragraphs to the issue, pointing out that many verbs in classes 7 and 8 (full reduplication) have unreduplicated counterparts in classes 1 and 3, and that the reduplication ‘adds pluractional semantics to the verb, i.e. the action is described as taking place several times’ (Kossmann 2004:61). Additionally, he observes that this pluractional meaning may develop into ‘doing sth. in a negligent manner’ (ibid.). Then finally he notes: ‘While semantics and derivational properties are regular with class VII, they are less regular with class VIII, which also includes many non-derived verbs with fuzzy semantics.’

What is ‘expressive’?

We’ll come back to the fuzzy semantics later. First, let’s set the stage by looking at that elusive term ‘expressive’. I would agree with Prasse that these words are expressive, using the word in the sense outlined in an earlier posting: words used by speakers to convey a vivid impression of a certain sensation or sensory perception. What are common properties of expressive vocabulary?6 Cross-linguistically, expressive words are often

  • phonologically marked, for example by deviant phonotactics and sometimes extra-systemic sounds
  • morphologically special, showing various types of reduplication and often a sense of open-ended variability
  • syntactically atypical, often being less well integrated into the sentence
  • semantically detailed, typically evoking an experience as a whole rather than picking out just one aspect of it
  • difficult to negate, because they are meant to evoke an experience rather than contradicting one
  • sound-symbolic, in the sense that closely related forms map onto closely related meanings (this is a type of diagrammatic iconicity)
  • multi-modal, in the sense that expressives more often than ‘normal’ words are accompanied by gestures

Note that since we are dealing with a function of language (the expressive function, Bühler/Jakobson) rather than with a word class per se, expressive vocabulary in Tuareg may or may not fit all of the above tendencies. So let’s ask ourselves how well our verbs fare as far as their expressivity is concerned.

Some questions

Are these verbs more marked phonologically than other verbs? Do they have atypical phonotactics? Do they feature sounds that are at the margin of, or fall outside, the phoneme inventory of the language?

Are they morphologically special? This is a definite yes, especially for the non-derived forms. Reduplicated roots are very atypical in Berber.

Is it easy to modify these words on the fly? For example, can we repeat the base (the part that is reduplicated) more than once to intensify the effect, or similarly, can we lengthen its syllables in an ‘analog’ way? Based on the ‘gradability’ of expressive vocabulary cross-linguistically, I would predict these verbs undergo such modifications more willingly than normal verbs. There we have a testable hypothesis.

How do these verbs cope with negation? Is it acceptable to use these verbs in the negative imperfective tense? If the preverbal negator wər/wær is used, does it negate just the verb or does it imply metalinguistic negation? Take the verb fərətfərət ‘sprout up here and there’. If we negate it, would we be saying “It is not sprouting up everywhere”, or rather “We wouldn’t use the word fərətfərət to describe this particular state of affairs”? The difference is subtle, but important. In the first case, we negate the state of affairs and we do so linguistically. In the second case, we are not negating anything at the linguistic level; we are rather passing a metalinguistic judgement concerning the suitability of the term fərətfərət in the given context. The prediction here is that cases of negated expressive verbs, if they occur at all, will tend to be interpreted as metalinguistic negation. If these verbs are indeed expressive, this makes perfect sense; for they are typically used to evoke an experience, not negate one.

Do we find sound-symbolic patterns in these verbs? A look at the examples given above does reveal some cases where closely related forms map onto closely related meanings (a type of diagrammatic iconicity). Take fələɣləɣ and fərəɣrəɣ for example: both evoke certain sounds, a soft and cristalline one and a thundering or clicking one respectively. The difference lies in the first consonant of the reduplicated syllable. If we add f̣ərəkrək ‘produce a small cracking sound’ to the picture, we see that we can also vary the final consonant of that syllable to evoke yet another kind of sound.7 Close inspection of a much larger inventory of expressive verbs would no doubt reveal more patterns like this. Perhaps some case could be made for phonesthemes, i.e. meaning bearing elements at the submorphemic level (compare English gl in glitter, glow, glare, glimmer, glimpse). More interesting would be templatic morphology of the type found in many languages that have large expressive inventories. Take for example the Siwu expressive template pVmbVlVV ‘fat+round’, in which vowel quality codes for size: pɔmbɔlɔɔ ‘huge, fat and round (as the stomach of a very fat man)’, pumbuluu ‘a bit fat and round (as a normal person’s stomach after having eaten well)’.

But I digress. Let’s ask two more questions, and then round up today’s posting.

Are these verbs frequently accompanied by gestures? This I don’t know at this point, but it is easy enough to investigate. The prediction here would be that they come together with gestures and facial expressions more often than other verbs. If so, there are a lot more interesting issues. (What kind of gestures? Can you leave out the verb and do the work with just the gesture?)

Last but not least, the ‘fuzzy semantics’ question. Are these verbs semantically detailed? You bet. Note that the data above is second-hand, lifted from a recent dictionary which is trustworthy, but not necessarily very elaborate in its definitions. Kossmann notes that it might be a good idea to look up these words in De Foucaulds grand dictionary (1951-52), which has much more semantic detail.8 Still, even the glosses given above are surprisingly specific, and anyone with some lexicographic experience will sense that these terms are probably quite hard to capture. Why would this be so? My tentative answer would be: these verbs are semantically detailed because their job is not so much to single out just one attribute (as adjectives or verbs of quality may do), but rather to evoke an experience as a whole.

Concluding remarks

If this posts brings up more questions than it answers, it is because these questions have needed asking for a while. If you feel there’s a lack of data in this post, you are right. I blame it partly on The Book Depository I ordered the two volume Prasse et al. dictionary (ISBN 9788772898445) with them, but they only sent me one volume, so I’m stuck now with half a dictionary. If you’ve come this far, thank you for your attention, and do feel free to weigh in with more questions or (even better) more data. (There is a place to leave comments below the footnotes.) Happy new year!


  1. Chaker, Salem. 1997. Quelques faits de grammaticalisation dans le système verbal berbère. Mémoires de la Société de Linguistique de Paris V:103-121.9
  2. De Foucauld, Charles. 1951. Dictionnaire touareg–français, dialecte de l’Ahaggar. Paris: Imprimerie nationale de France.
  3. El Mountassir, Abdallah. 2005. Lexique et perception de la realité en berbère. L’exemple du terme azaɣar. Faits de Langues 26:259-268.10
  4. Heath, Jeffrey. 2005. A Grammar of Tamashek (Tuareg of Mali). Berlin: Mouton de Gruyter.
  5. Kossmann, Maarten. 2004. [Untitled overview of Tuareg phonology, morphology, and syntax]. Ms., Leiden. (120p.)
  6. Prasse, Karl G., Ghoubeïd Alojaly, and Ghabouane Mohamed. 2003. Dictionnaire touareg-français. Copenhagen: Museum Tusculanum Press, University of Copenhagen.
  7. Sudlow, David. 2001. The Tamasheq of North-East Burkina Faso: Notes on Grammar and Syntax. R. Köppe Verlag.


  1. The great two volume dictionary by Prasse, Alojaly and Mohamed (2003) that is, not the grand four volume one by Père Charles de Foucauld (1951-2).
    By the way, a tip: the Prasse et al. 2003 dictionary is incredibly low-priced at €63. Note that this is for the two volumes together, so that’s a whopping 1031 pages in hardcover! Shipping and handling costs are a bit high (€22) if you’d order directly from Museum Tusculanum Press, the publisher, but it’s possible to find it in other places. []
  2. Tuareg verbs can be divided into a number of morphological classes. Different authors have come up with different ways to do this, and I’ll just follow Kossmann here (who himself mostly follows Prasse 1972-4) in distinguishing 19 major verb classes. The two parameters that aid in making the distinctions are: “(a) the formal make-up of the (consonantal) stem (number of consonants, geminations, reduplications, etc.), and (b) the morphological behaviour of these verbs as regards the aspectual apophony.” (Kossmann 2004:55ff.). []
  3. the prefix is a piece of reflexive morphology []
  4. the prefix is a piece of reflexive morphology []
  5. Similarly, Heath in his recent grammar of Tamashek lumps most of them together in a phonologically motivated class of ‘superheavy’ verbs but doesn’t say anything about their semantics (Heath 2005:102-3, 345). I haven’t had the chance to look up what Sudlow (2001) has to say on the issue. He does have a section on derived verbs (§3.18, p. 87ff.). []
  6. Many of the points in the overview below were developed in close collaboration with my colleague Sylvia Tufvesson, who is working on expressives in Semai, a Mon-Khmer language of the Malay peninsula. []
  7. Lest anyone points out that, say, fələzləz ‘bloom, blossom’ is close to these in form but not in meaning, I should note that it would be unreasonable to expect all of these words to be part of just one sound-symbolic field. []
  8. Unfortunately, I haven’t got those four volumes on my desk. Can anyone tell me whether the 2005 reprint by L’Harmattan is worth the deal? []
  9. This article by Salem Chaker touches upon expressivity in Tuareg. Chaker hypothesizes that geminates and long vowels, now widespread in Berber verbal systems, originated as a piece of expressive phonology that has been grammaticalized since. Not a word about reduplication or the huge number of expressive verbs, though. []
  10. More on this nice article by El Mountassir in a future posting. []

2 thoughts on “Expressive verbs in Tuareg”

  1. Great post – I hope to answer it with one on the Kwarandzie equivalent sometime. (One of these is actually present in K. – ffedfed “to thrash about” (normally used of a sheep just after its being slaughtered.) K. expressive verbs, like Arabic ones, normally consist of a reduplicated CeC – eg “be very fat”, bbetbet “be tiny”, lleghlegh “be full to point of nausea”, mmet.met. “suck hard, suck dry”, qqedqed “be spotted”. I’m not yet sure whether they constitute a unified class, or indeed how to treat them, but they often seem to have an intensifying or distributive sense (and CeC reduplication with pluractional/distributive meaning is regular for a hard-to-define subset of verbs.)

Leave a Reply

Your email address will not be published. Required fields are marked *