The definite article in both Brythonic and Goidelic is believed to derive from an unstressed form of the demonstrative pronoun *sindos (compare the development of English ‘the’ < OE. se ‘that’ and F. le, la < L. ille, illa ‘that’).  The Brythonic languages each have different forms of the definite article:

  • W. yr before vowels & h-, and before consonants,
  • C. an in all environments,
  • B. an before dentals, al before l- and ar elsewhere.

Clearly, there have been various degrees of assimilation in each of the languages, no doubt due to the unstressed nature of the word, and similar processes have occurred in the Goidelic languages: G. an becomes am before labials, and M. yn often becomes before consonants.

In Welsh, the form in -r became ubiquitous before the Old Welsh period (e.g. ir bis bichan ‘the little finger’, dir escip ‘to the bishops’). Traces of an article in -n still exist in the language, however: y naill ‘the one’ is for *yn aill (compare B. an eil, C. an eyl).


The evidence for the Cumbric article is limited, but there is reason to believe that the Cumbric system retained an article in -n beside one in -r, but perhaps also one without a final consonant:

Triermain (Treverman)
Terregles (Travereglys)
Tranent (Treuernent)
Traquair (Treverquyrd)*
Trailtrow (Travertrold)
Trailflat (Traverflet)
Trabrown (Treverbrun)
Trevercarcou (lost)
Trochrague (Trevercraig)
Trevergylt (lost)
Troneyhill (Treveronum)
Tallentire (Talentir)Penicuik (Penikok)
Penyghent (Penegent)
Treales (Treveles)

*Traquair also appears as Trevequor (c1153) and Travequayr (12C-13C)

Distribution of place names with the article with Vr in red, Vin green and V in blue.

To this list we might add Carluke (Cerneluke 1320), which is said to have Cu. *cair ‘fort’ as its first element and perhaps *luχ ‘loch’ as its second, so ne might reflect an article in -n.  But the derivation is very uncertain and may be Gaelic, so it is hardly a strong case.  Koch suggests that Liscard on the Wirral (Lisenecark 1260) is from Cu. *lis ǝn carreg ‘court of the stone’ with a nasal article, but the Wirral is well beyond what most authorities define as the Cumbric region (even Koch himself) and is only a stone’s throw from Wales.  To include it in this discussion on the basis of its having -n is entirely circular.

One particularly interesting example for this discussion is the lost place name Trevenlene (bef. 1153), which occurs twice in the Book of Kelso alongside Traverlen (four times, 1165-1214) and Treverlene (1245-54). Watson explains this as tref ‘wr lên (sic.) ‘village of the lector’, which seems rather forced and unlikely. The second element is probably the article, given the proliferation of examples with Cu. *treβ ǝr ‘village of the’. The final element is uncertain, the final -e suggests a long vowel, which tends to discount Cu. *linn ‘lake’ as a possibility. The equivalent of W. llên is possible, the word came to mean ‘scholars, priests, clerics’ in MW. and may have had similar meaning in the north. Other possibilities include Cu. *lein ‘strip of land’ (W. lleyn, llain) or, perhaps, Cu. *lejon ‘legion’ (W. lleon, lleng).


The evidence given here offers no straight-forward insights into the Cumbric article.  It is clear that an article of the type Vexisted and that it was the preferred form. But what are we to make of the evidence for the V  and Vn types?

Though the evidence for a vowel-only article is limited, it does seem to indicate that it was a genuine phenomenon and not merely an accident of recording or borrowing.

Support for the idea that Cumbric retained an article in -n is exceptionally scarce, but, the two examples are compelling: Tallentire is strong in itself since an alternative etymology is difficult to come by; and Trevenlene occurring alongside (and as an earlier form of) Treverlene seems self-evident, if we accept the former as genuine and not a scribal error for the latter.

We can therefore suggest, with varying degrees of certainty, that all three forms of the article did exist in Cumbric and remained in used until a relatively late stage of the language. However, there are no patterns in the evidence given here to suggest how the different forms were employed. The article used was apparently not chosen on the basis of the following sounds (as in modern Breton and Welsh), nor does it appear to be driven by morphology, such as the gender or number of the following noun.

Given the paucity of evidence, it is unlikely that we will ever unwrap the details of the Cumbric article fully. Names first recorded in the 12th and 13th centuries may have been borrowed at any time over the preceding centuries and diachronic changes have been obscured. It is possible that the names were borrowed at a time where an -n article was being lost and the vowel-only article introduced.