PIE Word Formation (3)

Accent movement

In the first approximation, the main principle of PIE accent movement is
quite simple: if a suffix or an inflectional ending containing a full
vowel is added to a base, it attracts the accent and the vocalism of the
base is reduced. Thus, for example, the root *kWer- 'make' extended with
the present participle suffix *-ént- becomes *kWrént- 'making', and the
gen.sg. of this, with the ending *-ós, is *kWrn.tós. In the simplest
case, represented by root nouns and stems with an accented suffix,
declension makes the accent move between adjacent syllables: the last
syllable of the stem (in the "strong" cases) and the inflectional ending
(in the "weak" cases). This type of accent movement is called
HYSTEROKINETIC:

nom. *p&2té:r *k^wo:n *dje:u-s
acc. *p&2tér-m. *k^won-m. *dje:-m (< *djeu-m)

but

gen. *p&2tr-ós *k^un-ós *diw-ós
dat. *p&2tr-éi *k^un-éi *diw-éi

Things are slightly more complex if the stem has the structure (a)
*CéC-C- or (b) *CéCC-C, *CéC-CC-. Here the accent may move between the
root vowel and the inflectional ending (the AMPHIKINETIC type),
especially in the stuctural type (a):

*pék^-u 'livestock', gen. *p&k^-w-ós (*& stands for a reduced vowel)
*krét-u-s 'strength', gen. *kr.t-w-ós
*d(H)ég^H-o:m 'earth', gen. dg^H-m-ós

(NOTE: In the subtype represented by *d(H)ég^Ho:m the *o of the second
syllable comes from a pre-PIE weak vowel which was retained and
lengthened in the nom.sg. of animate nouns and in collectives. The
endings that caused the lengthening were lost in some environments.)

It seems that in the more complex type (b) the expected weak cases such
as the genitive *C(&)CCC-ós received an epenthetic full vowel to break
up an overlong stem-final sequence of consonants. The result was
*C(&)CCVC-ós, and the V (usually = *e, though some branches have *o if
the stem-final consonant is *w) drew the accent from the ending,
yielding *C(&)CCéC-s in the so-called PROTEROKINETIC type:

*pért-u-s 'ford', gen. *pr.t-óu-s ~ -éu-s
*gWHér-mn. 'heat, gen. *gWHr.-mén-s
*páh2-wr. 'fire', gen. *ph2-wén-s
*h1éi-tr. 'route', gen. *h1i-tén-s

The original dependence of amphikinetic and proterokinetic accent
movements on the structure of the stem was formulated by Oswald
Szemerényi. It must be noted, however, that their actually observed
distribution is far from neat, since analogical cross-influence led to a
partial confusion between the two types already in PIE. In many cases
proterokinetic nouns developed amphikinetic forms such as *gWHr.-mn-ós
beside *gWHr.-mén-s. There was also a strong tendency to "immobilise"
the accent in some common types of nouns, in particular thematic stems
like *wl.'kWo-, in which the same syllable is accented throughout the
paradigm, or the neuter *-es- stems like *k^léw-os, gen. *k^léw-es-os
(with some archaic forms demonstrating that the accent was once mobile).

The vocalism of originally mobile types was also levelled out,
especially in old proterokinetic nouns with their complex pattern of
vowel alternations (the accent moving between two vowel slots in the
same stem). From *pér-tu-s/*pr.t-éu-s we get a regularised paradigm:
nom.sg. *pr.-tú-s, gen.sg. *pr.-téu-s; similarly, from
*mén-ti-s/*mn.-téi-s 'mind, thought' we get *mn.-tí-s/*mn.-téi-s, with
the accent immobilised and vowel alternations restricted to the suffixal
syllable.

The _old_ static type (as opposed to later types with secondarily
immobilised accentuation) will be treated separately in the next part of
my presentation.

Piotr