Very little evidence for borrowing:

a word has Arabic phonological and/or morphological features, but no source could be established. This may either mean lack of documentation (but Maghribine Arabic boasts very good dictionaries), or an expressive derivation from some non-identified item.

Perhaps borrowed:

I do not believe the proposed etymology, but I cannot rule it out altogether.

Probably borrowed:

I suppose it is a borrowing, but feel there are problems.

The field “Integration” has only been filled in for nouns, and refers to types of morphological integration. “Highly integrated” refers to borrowed nouns with Berber morphology, “Intermediate” to those nouns which have quasi-Arabic morphology, and “Unintegrated” to those which have faithful-Arabic or European morphology.

Phonological criteria are left aside, as in Tarifiyt phonological integration is somewhat more complicated than elsewhere, while morphological integration does not always follow the same lines as phonology (so you may have a recent European loanword which is hardly integrated phonologically, but which is entirely integrated into Berber morphology, such as a spanyu “Spanish”, pl. i-spunya, which has the recent loan phoneme p, but full Berber plural Ablaut).

The degree of phonological integration can be easily derived from the “period” and the “donor language” field:

– full integration: pre-Islamic loans and early Islamic loans

– partial integration (i.e. introduction of Arabic loan phonemes, but application of a number of Tarifiyt Berber phonological rules, such as rhotacism and vocalization of /r/): Islamic period, partly also “after 1900”.

– hardly any integration (only the almost automatic weakening of stops): “Islamic period” + “Donor language Classical Arabic” or “Donor language Standard Arabic” + partly “after 1900”.

My hunch is that full phonological integration is only found in loans from before about 1000 CE, that partial integration is found in loans between 1000 CE and 2007 CE, while hardly any integration is found in loans from 1920 CE to 2007 CE. But this I can hardly substantiate.

The transcription system has been adapted to the technical and scientific context of the database. The following conventions have been used:

First column: My WOLD usage

Second column: IPA

Third column: Standard linguistic transcription of Tarifiyt

Fourth column: Official Tarifiyt orthography (Lafkioui 1997)

ā a: ā ar, er, re

ā’ a:ˁ ạ̄ aṛ, eṛ, ṛe

ƀ β ḇ b

č ʧ č lt

ç ɕ ḵ k, y

d’ dˁ ḍ ḍ

đ δ ḏ d

đ’ δˁ ḏ̣ ḍ

ə ə e e

ǧǧ ʤ(:) ǧǧ ll

ħ ħ ḥ ḥ

ia ea ia ir

l’ lˁ ḷ ḷ

r ɾ r r

r’ ɾˁ ṛ ṛ

ř r ř l

š ʃ š c, k

s’ sˁ ṣ ṣ

ŧ θ ṯ ṯ

t’ tˁ ṭ ṭ

ua oa ua ur

y j y y

z’ zˁ ẓ ẓ

ž ʒ ž j

ʔ ʔ ’ ’

ʕ ʕ ɛ ɛ

Lafkioui, Mena. 1997. Propositions pour la notation usuelle à base latine du rifain. Paris: INALCO.

In this field, simple nouns have been treated as unanalyzable. As the great majority of nouns have in fact an analyzable morphological structure, it would have been possible to make a different decision.

In the Age field, I do not go beyond Common Berber; reconstructions of earlier levels (Proto-Afroasiatic) have not been taken into account. The assignment of ages to non-borrowed words has been done intuitively; if I happened to know that the word is well-attested in Berber, it was assigned to the Common Berber age (which means: possibly reconstructible into Proto-Berber), otherwise I chose "No Information". In cases where the non-borrowed word contains loan phonemes from Arabic, as is usual in expressive formations, it was of course possible to date the word in its current word form to a period when Arabic was a factor in the Maghreb, and the age "Islamic Period" was assigned.