Deixis and anaphora in Berber: the case of din and of the pair dd/nn

by Fernand Bentolila

Linguists often oppose anaphora with deixis. According to Francis Cornish (Anaphoric Relations, p. 3),


« Deixis proper involves the introduction of a new object

of focus within some universe of discourse, while anaphora

presupposes a common, pre-existing focus. »


However in many languages the deictic forms are also used as anaphora. Such is the case in Berber of din and of the pair dd/nn. After placing din in the deictic paradigm to which it belongs, we will study its anaphoric uses and will discuss the grammaticalisation of din as a relative. Then we will deal with the approaching and moving away particles dd and nn, demonstrating that in certain contexts these deictic particles fonction as anaphors. All our examples will be taken from the Berber dialect of Aït Sehrouchen (Central Marocco).


There is in Berber a class of grammatical determiners of nouns containing six members :


kull : « every, each »                          man : « which »

flan-flani : « such »                            –u : « this »

din : « that »                                      –inn : « that… there »


These units exclude one another in a string of elements, that is to say that a noun cannot be determined, for instance, by kull and din at the same time. Amongst these determiners there are three deictic forms : –u designates an object close to the speaker, –din an object close to the addressee and –inn an object that is neither close to the speaker nor to the addressee. There is therefore in this dialect a micro-systeme with three terms comparable to those of Latin (cf hic, iste, ille), Spanish (cf este, ese, aquel) or Korean (i-, kt-, ca). The fundamental opposition is that between –u (first person) and –inn (third person). The deictic sense of din (second person) is precarious and overall more difficult to establish ; however it remains used. In particular if the addressee were some distance from me and if he were holding a pen in his hand, I should say to him :

i Stilu din : « give me this pen ».

u and –inn would be impossible in this example.

There is in a tale an example where this deictic meaning of din can be found (arrow pointing to the addressee’s sphere) : Hammou asks someone to tie him to a tree. A while later a new protagonist arrives. Hammou tells him that he is tied to the tree because it is a way to cure rhumatism. And the reply is :

džu Sžṛt din asfar, aS i « if this tree / (to which you are tied) is a remedy, tie me up ».

But din can have a second meaning which is much more frequent than the previous one : it is the anaphorical interpretation. In effect, in most cases din is used to refer to an antecedent mentionned in the preceeding context or belounging to the discourse environment, and therefore present to the mind of both the speaker and the addressee : « the man in question, the object about which one is talking ». We find there two meanings again (deictic and anaphora) in the composed pronouns : udin « that one », as opposed to wu « this one » and to winn « that one there », and adin « that » as opposed to ayu « this » and to ayinn « that… there ».

For the adverbs of place there are only two deictic forms : da (here) and dinn (there). Anaphora is expressed by di-s, litteraly : «  in it », that is to say « in the place in question ».

In conclusion we can say that Berber uses a deictic form for anaphoric uses. If we compare this with Latin, there seem at first to be differences since Latin has at its disposition, along with its three terms micro-systeme (hic, iste, ille), a pure anaphore : is. But, as Alfred Ernout and François Thomas note :

« La répartition précédente ne dura qu’un temps, et, au      cours du latin, ces différents termes se sont de plus en plus employés les uns pour les autres. Is sortit d’usage assez vite (…) Hic s’affaiblissait et contribuait à remplacer is (…) Ille encore plus que hic tendait à se substituer à is comme anaphorique (…) Iste cessait d’être lié à la deuxième personne. (Il se maintenait) empiétant parfois sur is et surtout sur hic ». (1)

(The previous repartition only lasts a certain amount of time and throughout the evolution of Latin these different terms are used more and more the one for the others. Is became redundant quite quickly (…) Hic became weaker and started to replace is (…) Ille even more so than hic tended to substitute itself to is as an anaphor (…) Iste ceased to be linked to the second person, maintaining itself by encroaching occasionally on is and especially on hic.)


We note therefore in Latin, as in Berber, an instability of the deictic form of the second person and the use of deictic forms as anaphora.