Dentalization of velar stops

Another non-standard phonological feature of Basque is dentalization of velar stops¹. This phenomenon is very often accompanied by expressive palatization²:

kakur 'dog' > ttattur
-ko 'diminutive suffix' > -t(t)o
konkor, kunkur 'hunchback(ed); bump' > ttonttor, t(t)unt(t)ur.

Dentalized variants of Latin and Romance borrowings like *fongo 'mushroom' > onddo, kipula 'onion' > tipula, (k)upa 'barrel' > dupa show the process was still going on in the High Middle Ages.

Unlike assibilation, dentalization only affects velars and can occur at any syllable onset and not only at word-initial. Sometimes, we can also find non-dentalized forms in Iberian corresponding to dentalized ones in Basque:

idi 'ox'³ ~ Iberian biki⁴ (PNC *bHe:mttɬɨ (˜ -u,-i)  'deer, mountain goat')

borda 'hut' (PNC *borGwV: (˜ -ǝ-) 'stall; shed, tower')
gurdi, burdi 'cart' ~ Iberian oŕke (PNC *k’wVrk’V 'something round or rotating')
idor 'dry'⁵ ~ Iberian bekoŕ⁶ (PNC *=iGGwAr 'dry, to dry')
urde 'pig' ~ Iberian uŕke (PNC *wHa:rttɬ’wǝ 'boar, pig')

Dentalization also arises in compound variants⁷ where a velar stop gets to final position:
begi 'eye' > *bet-
behi 'cow' (EPBasq. *pek^hi) > *bet-
ogi 'bread' > *ot-
zohi 'clod of earth' (EPBasq. *tok^hi) > *zot-
______________

¹ Commonly found in children's language.

² As in the case of assibilation (see), this is why Vascologists haven't indentified it.

³ Also non-standard pit(t)ika 'goat kid' and similar forms.

⁴ Compare also Spanish pequeño/a (without palatization!) and Italian piccino/a 'little' with Catalan petit/a and French petit/e 'id.'

⁵ There's also ador (L), a form contaminated by agor 'dry'.

⁶ There's also Basque pegor 'sterile, poor'.

⁷ Called combinatory forms by Vascologists.

Assibilation of initial stops

The lamino-alveolar sibilant [s] <z> isn't a native consonant in Basque at word-initial, its possible sources being these ones:

1) The retention of an etymological sibilant in non-standard varieties vs. its loss in standard Basque, as in zapo 'toad' (a loanword from Semitic *ɬ’abb- 'a kind of lizard') vs. standard apo.

2) The result of Iberian *t- in loanwords. We've already seen that in a previous post.

3) The assibilation of an initial stop, either velar (most frequent) or labial (very rare), which often undergone expressive palatalization¹ into [ʃ] <x> or [tʃ] <tx>. For example, non-standard kakur 'dog' has given the assibilated forms zakur, txakur, xakur, and Romance *capel 'hat' has given txapel, xapel, zapel.

An example of assibilation of a labial stop is zegi 'milky cow' vs. behi 'cow' < EPBsq. *pek:i.

The coalescence of assibilation and expressive palatization in the same  word has made Vascologists² incapable of identifying the former. In Basque only coronal consonants [s, ś, ts, tś, d, t, n, l] (and in some dialects also [r, ŕ]) can undergone palatalization, so in order for a velar to be palatalized, it must be first converted to a sibilant.

The chronology of assibilation is ancient, as it's already found in Iberian. For example, zaldi 'horse' (Iberian saldu) is related to Germanic *kulta- 'colt'³, and zamar 'sheepskin jacket'⁴ is related to Germanic *xamísa- 'clothes, skirt'⁵ (PNC *χ^ʕa:mV ‘skin, cloth’).

____________
¹ A common linguistic device in Basque used to denote diminutive or affective meanings.
² Bengtson among them.
³ Nikolayev reconstructs IE *g^(w)Ald- 'foal, young of an ass', related to NEC *gwalV 'horse'.
⁴ A non-standard word borrowed into Romance (Spanish zamarra, Catalan samarra).
⁵ Hence Spanish camisa 'shirt'.