No CrossRef data available.
Published online by Cambridge University Press: 01 September 2010
The standard view that thematic verba pura contained Proto-Germanic j is examined in detail and rejected: we must reconstruct PG +sē1e/a- +rōe/a-, not +sē1je/a- +rōje/a-. The Gmc. verba pura typically have stems ending in ē1 (Go. saian, OE sāwan, ON sā, OHG sāen ‘sow’) or ō (OE rōwan, ON rōa ‘row’). While ō in these verbs seems always to reflect PIE o + a non-coloring laryngeal E, ē1 occurs in verbs with laryngeals of any color: saian <+seE-, OHG krāen ‘crow’ <+greA- (a-coloring), OE cnāwan ‘know’< +ǵneO- (o-coloring). A solution is offered according to which in Germanic, E was lost relatively early, e and o in hiatus were then lengthened to ē1ō, after which remaining o > a. Later, when A and O were lost, a in hiatus was lengthened to ā and reanalyzed as /ē1/. However, +knē1- might instead reflect a Narten present in +ǵnēO-.*