Abstrakt
In recent Indo-Europeanist literature, it is common to encounter statements concerning phonological feature [voice] and PIE laryngeals in a wide spectrum from explicit commitment to specific feature- segment combinations to reserved observations on possibilities and probabilities. This article aims to consider some implications of these statements, the possibilities to prove or disprove any such claims through distributional patterns of the relevant segments.
Namely, I consider their presence in consonant clusters that should show traces of voicing assimilations if laryngeals were phonologically +/-[voiced]. I will argue that the implications involved in a strong commitment to the [voice] as a phonological feature in PIE laryngeals is not warranted by data, which may either mean that the feature was not phonologically relevant, or cannot be shown with any certainty to be the property of any of the three segments in either of its values, +[voiced] or -[voiced].
This is in principle reconcilable with laryngeals-qua-fricatives in as much as they are reconstructed as having sonorant allophones in consonant clusters, for which there is typological support from Cairene Arabic.