Hallo conlangers!

As some of you already know, I am very interested in the languages of
prehistoric Europe, and try to "re-create" some of them in my conlangs.
After a longer hiatus, I have picked up research in these matters again
a few weeks ago. Of course, if you do both research in Paleo-European
linguistics and make your own conlangs based on that research, you
always must know which hat you are wearing at the moment, and not let
your research be led astray by using "evidence" from your conlangs!

By their nature, these languages are hard to know about. Yet, they are
not an entirely unknown entity as there are ways of looking into the
past, such as the celebrated comparative method by which such lost
common ancestors such as Proto-Indo-European can be reconstructed,
internal reconstruction which looks even deeper starting with
irregularities which show traces of old, formerly regular patterns, or
the analysis of substratum loanwords.

First, let me begin with a round-up of the pre-IE languages of Europe we
know of.

1. Living languages: in western Europe, only Basque. On the southeastern
periphery of Europe, there are the three Caucasian families. Uralic is
not older than IE, but spread in parallel to IE. (It cannot safely be
ruled out that Basque is a latecomer, but so far there is no evidence of
that.)

2. Extinct languages attested in writing: there are plenty of these in
the Mediterranean zone. On the Iberian Peninsula, we have Iberian, which
shows a few resemblances to Basque which, however, may be deceptive, and
Tartessian, which looks nothing like Basque or Iberian. A few scholars
conjecture that Tartessian was Celtic, but most scholars remain
unconvinced (see note 1 below). Both Iberian and Tartessian are so
fragmentarily attested that we can hardly say anything about them. In
Italy, there's Etruscan, which, however, may constitute a Late Bronze
Age (ca. 1200 BC) immigration from Anatolia (see note 2 below). The
sparsely attested Rhaetic language in the central Alps seems to be
related. In the Aegean, Lemnian, known from two inscriptions, is
obviously related to Etruscan. Minoan on Crete is written in an older
hieroglyphic script (undeciphered) and a younger script called Linear A
(decipherment under way); the change of script may have been motivated
by a structural change in the language. The much younger Eteocretan
language, attested in a few fragments written in the Greek alphabet, may
be a descendant of Minoan, but may just as well be a language that came
to Crete after the fall of the Minoan civilization. The situation on
Cyprus is similar to that on Crete. There is an unknown language in the
Cypro-Minoan script (so named because it appears to be an offshot of the
Minoan scripts, which of course says nothing about a relationship
between the *languages*), and a much later unknown language,
Eteocypriot, in the Cypriot syllabary (deciphered, and otherwise used to
write Greek). Further north, there are the enigmatic Vinča symbols in
the Lower Danube area, of which we don't know whether they were writing
at all (if they were, they'd be the oldest known writing, as they are
about 2,000 years older than the oldest Sumerian writing!). There is
otherwise nothing in this category north of the Alps, with the possible
exception of Pictish, which, however, is now considered Celtic by the
majority of the relevant scholars.

3. Traces of unattested languages: Linguists have repeatedly tried to
recover lost languages from loanwords, geographical names and structural
changes in known languages. This is to the largest part uncharted
territory, and controversies abound.  A famous example is the substratum
some scholars suspect behind the aberrant development of the Insular
Celtic languages. This has often been held to be related to Semitic, but
VSO word order and some typological correlates thereof do not prove a
Semitic connection. More generally, the development of the Late PIE
tripartite aspect system to a single past tense in northwestern
Indo-European languages may be due to a substratum that did have a
simple present/past system rather than the more intricate Late PIE
tense/aspect system. Also, these languages contain a number of words
without good PIE etymology, as does Greek (or any other IE language). A
much-discussed trace of an unknown language family is the "Old European
Hydronymy" (OEH), a network of seemingly recurring river names spanning
much of Europe; this has been ascribed to PIE (Schmid), an unknown IE
branch (Villar, Anreiter et al.), a "Vasconic" language family with
Basque as the last surviving member (Vennemann), and others. See also
note 3 below.

Note 1: The Tartessian inscriptions in the southwest of the Iberian
Peninsula are the oldest inscriptions of the entire peninsula, dated to
the 7th century BC. The Celtic hypothesis (John T. Koch) seems to make
sense as there is no shortage of Celtic names in the area (such as the
famous king Arganthonis), but requires an appropriately early arrival of
Celtic in the Iberian Peninsula. Koch's interpretation of the Tartessian
inscriptions as Celtic has been reviewed negatively by many scholars. In
my opinion, it is more likely that the Tartessian inscriptions represent
a pre-Celtic language, and that the Celticization of the region marks
the *end* of Tartessian literacy.

Note 2: Proto-Tyrrhenian, the common ancestor of Etruscan, Rhaetic and
Lemnian, may have been the language of Troy. The Roman foundation myth
may point at a Trojan origin of the Etruscan, perhaps as part of the
"Sea Peoples" migrations mentioned in Egyptian chronicles (in
pre-republican times, Rome was ruled by an Etruscan aristocracy). Some
Greek historians also entertained such a notion, and the Lemnian
language may represent a descendant that stayed closer to home. However,
Woudhuizen's hypothesis that Etruscan descends from Luwian is wrong;
Luwian is related to Hittite and thus an IE langauge.

Note 3: I suspect a sister language family of IE here which I call
"Aquan". This language family would have spread in the Neolithic and
left traces in the Old European Hydronymy and in the form of substratum
loanwords in western IE languages, chiefly Celtic and Germanic.
According to my ongoing research, this language branched off before PIE
developed ablaut, had a three-vowel system */a i u/, was of the
suffixing agglutinating type, and had a simple present/past verb tense
system like Hittite. The peculiarities of Insular Celtic may be due to
the Aquan language of the British Isles.

--
... brought to you by the Weeping Elf
http://www.joerg-rhiemeier.de/Conlang/index.html
"Bêsel asa Éam, a Éam atha cvanthal a cvanth atha Éamal." - SiM 1:1