the Wheel and Dating PIE or NW-IE

[Robert Orr]

[ Moderator's note:
  The following posting is edited from two versions submitted by Mr. Colkitto,
  each with some problems.  I am taking the liberty of posting this edited
  version so that the discussion can continue.
  --rma ]

>> Could someone spell out in more detail what these 'archaic characteristics'
>> are? Or at least the most salient ones. >>

> -- with respect to the Baltic languages, particularly the declension of the
> noun and adjective, with seven cases, singular and plural, and preseveration
> of the dual as well in some dialects.  This is almost completely unchanged
> from PIE.

> The Baltic verbal system is more innovative, but the PIE present tense is
> well-preserved, and the future represents a PIE disiderative formation
> revalued as a simple future.

> There PIE lexicon is preserved to a really startling degree:  eg, *dubus >
> dubus (deep), *gwous > guovs (cow), *h(1)rudh > rudas (red), etc.

> The most archaic Baltic languages (Lithuanian and Old Prussian) are about as
> close to PIE as some of the very first attested IE languages.  Conservatism
> of a mind-boggling degree.

This argument has already been played out in Baltic and Slavic studies.  The
term "archaic" in linguistics should be used with extreme caution.  Once again,
I am taking the liberty of quoting myself (pp. 91-94)

3.7 Slavic and Baltic: Relative Degrees of `Archaism'

Although at first sight Baltic appears far more `archaic' than Slavic; in fact,
certain scholars have even advocated using Baltic, especially Lithuanian, data
as a check on CS reconstructions (see especially Otkupšcikov 1974, 1983,
Jasanoff 1983a: 142, although for criticism of this view see, e.g., Stankiewicz
1988), this generalisation needs some clarifying in at least two respects.

3.7.1 Defining `Archaic'

In historical linguistics (and probably in other related fields) the terms
`more archaic' and `less archaic' need some quantifying.  To take one fairly
obvious, albeit extreme, example, probably no scholars would rate English as
`more archaic' than German overall, and yet for certain features, in phonology,
morphology, and syntax, this is an fair statement, e.g.,

1) The preservation of Gc þ in English as opposed to its loss in German,
e. g., E three - Ge drei < CG *þrij-.

2) The related preservation of unshifted Gc t, d in English as opposed to their
development to c (orthographic z), t under the Second Sound Shift in German,
e.g., E ten - Ge zehn (< CG *tehun).

3) The preservation of IE w in English as opposed to its development to v in
German, e.g., E win - Ge (ge)winnen [(ge)vinen].

4) The preservation of a three-fold alternation in certain Gc Class I strong
verb paradigms, e.g., E drive (< IE *-ei-) - drove (< IE *-oi-) - driven (< IE
*-i-), as opposed to Ge treiben (< IE *-ei-) - trieb - getrieben (the
generalisation of the reflex of CG *-i-).

5) The failure of English to develop HAVE-constructions such as Ge Ich habe
Hunger `I have hunger' "I am hungry".

6) The preservation of 1st sg am in English as opposed to its loss in German.

7) The preservation of certain lexical items from Common Germanic (and IE) in
English as opposed to their loss in German, e.g., E tree, choose, ask.

Of course, it would be very easy to think of far more examples where German
preserves archaisms where English has innovated, especially when dialectal
materials are considered; the point here is that `more archaic' and `less
archaic' are relative terms, to be used with a degree of caution.  Within
Slavic, Russian may be seen as `more archaic' than Polish with regard to
certain syntactic constructions and lexical items (Orr 1992), whereas from the
point of view of the evolution of gender Polish may be said to be more archaic
in some respects than Russian (cf. Slobin 1985: 1194, 1216; Weiss 1993:
99-101). Manczak 1991: 74-76 makes a similar proposal regarding the oft-cited
status of Sardinian as the most archaic Rmc language.  This perception of
Sardinian is based on the fact that it preserves the original Italian velars
before front vowels; Manczak, however, shows that from the point of view of the
lexicon, standard Italian is actually more archaic than Sardinian.

3.7.2 Defining `Baltic'

`Common Slavic', `West Baltic', and `East Baltic' are relatively
straightforward concepts; `Balto-Slavic' is much less so; but the problems in
reconstructing `Common Baltic' are immense.  In fact, it has even been
suggested that Slavic is originally a type of `West Baltic' (Zeps 1984), cf.
also Ivanov and Toporov 1961; Toporov 1988.

At this stage it might be advisable to summarise the points where Slavic may be
said to be `more archaic' than Baltic in the nominal declension.  In most of
them it will be noted that Slavic disagrees with only a part of Baltic in many
forms.  Some will necessarily be contentious, cf. especially (6), (7), (8).

1) The preservation of the neut in Slavic as opposed to its loss over most of
Baltic; in this context it is startling, that Lithuanian, which is usually
taken as the most `archaic' form of Baltic, has almost completely lost the
neut, while Old Prussian has preserved it (but see Schmalstieg 1992).

2) The related restoration of the IE anim/inanim distinction (see 3.2.3, and
Kry ko 1994: 198).

3) The preservation of a neut *s-stem paradigm in Slavic as opposed to its
near-loss in Baltic, e.g., OCS slovo `word' - gen sg slovese.

4) The preservation of a * -stem paradigm in Slavic as opposed to its loss in
Baltic, e.g., OCS svekry `mother-in-law'- gen sg svekru ve;

5) The more recent survival of length in the nom sg form of the fem * -stems in
Slavic than in Baltic, cf. OCS raka `hand'< *rank ; Lith rankà `hand' <
*ranka would have developed into *ranko without shortening, and we might refer
to the fem *a-stems in Lithuanian, as in Germanic.  It should be noted,
however, that here Old Prussian agrees with Slavic against East Baltic, cf.
Lith galvà `head'; Latvian galva id < *- as opposed to OPr galv id; OCS glava
id < *- ;

6) The failure of Slavic to develop the postposition *-en as an affix, cf. OCS
syn x ; Lith s nuosè `son' loc pl, Old Lith -su.

7) The development of a gen pl form in Baltic which may reflect *- m as opposed
to the lack of any such form in Slavic.

8) The greater extension of *-m as an acc marker in Baltic than in Slavic.

9) Finally, one might also cite the fact that while extra cases have emerged in
both Baltic and Slavic, Lithuanian might be said to have carried this
development further, see Stang 1966: 228-32, e.g., galvà `head'; acc sg
gálv , illative sg galvõn, gen sg galvõs, allative sg galvõsp, loc sg
galvojè adessive sg galváip; within Slavic, Russian might be said to have
developed new cases in the Gen-2 and Loc-2, cf. Jakobson 1932, 1958.  In any
event, this point only serves to highlight the difficulty of categorising whole
groups of languages as `more or less archaic'; one part of Baltic (Latvian) may
have lost cases (or failed to develop extra ones).

(references available off lst)

Robert Orr