I was pretty impressed with Ruhlen's book, but one of the criticisms of his method troubled me. It is charged that his samples across language families stretch the semantics, using here a word for "mouth" and there a word for "talk" and elsewhere a word for "tongue," etc. The lack of rigor in his method allows him to paint a misleadingly convincing picture.
The same sort of semantic spread is seen within Indo-European, for example:
In post 103, there are clear IE cognate words that mean "vulgar term for female private parts (or the moisture therein)" in English, and "whore" in French or Spanish.
Or consider
via (Latin) = "road",
je vais, tu vas,.. (French) + "I go", You go"...
and English "wagon"
These are all cognates
Another example
agricola (Latin) = "farmer"
acre (English) (unit of measuring farmland)
And how about decem (Latin)= "ten"
Index (Latin) = "forefinger"
digit (Latin) = "finger" all of these have similar sounds (G and c are always hard in Latin), and are cognate to English "toe"
And one final one, which Ruhlen et al posit is universal: Gynos (Greek) = "woman" (as in gynecology)
cognate to English "queen" and probably cognate to English "c*nt"