Periphrastic tenses

According to Trask "The History of Basque" that language has
periphastic constructions both for the past ('be'/'have' plus
participle in the style of Germanic and Romance) and for progressive
tenses (copula plus a particple in a locative construction, in the
style of Italian, Spanish, Portuguese, English, Gaelic, Dutch, Danish
(but not Swedish or Norwegian afaIk) and Finnish; possibly (North?)
German <er war beim Schreiben> or, as I think I've heard, <... am
Schreiben>, but it's not commonly used).

Trask mentions that several linguists think this is a calque from
Romance in Basque. Vennemann mentions that Simon (1939) has argued
for a loan in the other direction; Vennemann, I believe, would opt
for a 'Vasconic' substrate.

Trask argues that the progressive must be recent in Basque, thus a
calque from Romance, since the gerund (imperfective participle) which
is used in the locative case (<-n>) in the progressive tense
construction in Basque exists in several forms in the dialects, using
different suffixes, the most common of which being <-te>, and
therefore the imperfective participle must be recent. Trask mentions
himself that in the past the pure stem form of the verb might have
served as a participle. If we assume that was the case, verbs the
stems of which ended in <-n> would have caused phonotactical problems
(the <-n-> simply disappearing, according to the historical laws of
Basque phology), when the locative suffix <-en> was added. Trask
describes a similar case where the perfective participle ending <-i>
complicated the phonotactics of verbs having stems ending in <-n> to
the point where it took a Trask to disentangle it (thus it certainly
was not transparent to the average speaker). Under those
circumstances it is to be expected that the language would
disambiguate the situation by using a new suffix more resistent to
the prevailing phonological laws.

In other words, Trask's argument that the periphrastic progressive
tense in Basque is recent, because the imperfective participle
(gerund) is recent, doesn't hold, and the calqueing might in
principle have been done from Basque (Vasconic) instead of into it.

Torsten