Think of modern people in Greece. If Greek suddenly "boomed" and had a huge demographic and linguistic expansion, the spread of Greek would be correlated with the expansion of several haplogroups (J2, E-V13, R1b, R1a etc.), and it's probable that in some places one of those haplogroups would randomly become much more common than the others, creating several Greek groups with different Y-DNA makeup, but initially very similar autosomally, and after some time, as they mixed with other local populations of the places where they had migrated to, they'd become also genetically different people;
Think of modern people in Greece. If Greek suddenly "boomed" and had a huge demographic and linguistic expansion, the spread of Greek would be correlated with the expansion of several haplogroups (J2, E-V13, R1b, R1a etc.), and it's probable that in some places one of those haplogroups would randomly become much more common than the others, creating several Greek groups with different Y-DNA makeup, but initially very similar autosomally, and after some time, as they mixed with other local populations of the places where they had migrated to, they'd become also genetically different people.
Ygorcs, thank you for your experienced analysis, it helps me to understand many things
The "founder principle", the loss of genetic variety created by the spread of a small group belonging before to a bigger group often causes hasardous and surprising consequences, as the very good exemple you gave us (N and I spreading a language commonly linked with R1a populations, etc.).
Meanwhile the study of haplogroups and the tentative to connect their history to the history of languages, remains interesting.
Y-Dna says nothing about language, physical nor social specifications of a human group, that's right. But Y-Dna allows us to study migrations of some of ou ancestry without writed sources, and this is a revolution in historiography.