There's an obvious similarity between the Proto-Indo-European reconstructed roots *leg- and *les-, both "to collect, gather", reflected in logos, Latin lego and German lesen respectively. I have not yet seen a comparison of the roots, and did not research explicitly, other than looking up Wiktionary.

I am hoping someone knows a reference from the top of their head.

I see three possibilities:

  1. There is a common root to the two reconstructions and a grammatical explanation for the dissimalirity.

    1.a) lesen did not have the connotation of reading. The sense is a phono-semantic loan from a christian incurser.

    1.b) "lesen" had the connotation to pick up hints for long.

  2. Both roots were the same but at different stages; Which would imply sound change not shared by both branches as early as in a PIE dialect scenario or at the latest in a scenario in which noble savage learns to read from forreigners--similar to 1.a)--and borrows the whole word mutating it slightly (Galician lesen, 3rd p. pl. subjunctive participle of ler "to read" < legere, reminds of Ger belesen "learned")

I'm sure that there is no main-stream hypothesis, otherwise I'd expect Wiktionary to note that. I'm pretty sure there are at best non-comittal opinions or unacceptable fringe theories at worst about questions concerning the timeline of PIE and unproven sound-changes. I think it is impossible to give an acceptable negative answer

I'd be surprised if a tentative synchronic explanation has been given on the example of exactly these two roots. A general grammatic theory to synthesize an answer would probably be beyond me; I expect I will not find the answer in introductory text-books (if that constitutes lack of prior research, the question should be closed, or put on hold for a long while).

So i guess I'm asking: Have I missed another principle possibility?

This could be deemed to broad and unlikely to result in a high quality answer and more likely to result in a low effort answer about the uncertainties of etymology, which would be yet too much effort to ask, for what it's worth, on the basis of the question.

What I really want to know is actually:

Is there a theory for a root *le-?-? Ideally, that should explain a few other *l- initial roots as well, perhaps even glossa.

  • A root *le- is impossible given PIE root structure (see e.g. linguistics.stackexchange.com/a/7272/2411). All the two roots have in common is initial l-, so the similarity is probably coincidental.
    – TKR
    Aug 2, 2019 at 0:02
  • 1
    It may be coincidence, or the two might have some relationship in pre-PIE times—there's no way to really say one way or the other, since the comparative method can't go back any farther than PIE (unless some exciting new evidence appears).
    – Draconis
    Aug 2, 2019 at 5:52
  • @TRK, that's why I wrote "*le-?-". Also, *so, *de, *ye etc exist. Suppose an aorist *l-H and the reflex would be *l- in most branches. Maybe "root" is not the correct etymology, sure.
    – vectory
    Aug 2, 2019 at 6:02
  • Sorry, I misunderstood the question mark. (But note that *so etc. aren't verbal roots.)
    – TKR
    Aug 2, 2019 at 16:41

1 Answer 1


At the time when Proto-Indogermanic was spoken, there wasn't any written form of it. It is therefore quite clear, that it also hadn't a word with the modern meaning of "to read". German usage like Kartoffeln lesen or Trauben lesen still preserve the old meaning "to collect".

The root form of Latin lego and many other words is *leǵ with a ǵ that becomes palatalized in the Satem languages as shown by an Iranian cognate. The synonymity between *les and *leǵ is indeed strange, but I am not aware of a deeper explanation. My first guess would be that in Germanic we have for some reason applied a partial satemization to that root or borrowed it from a Satem language at an early stage in its history, but the existence of a Hittite (li-iš-ša-an-zi) cognate to lesen makes this hypothesis less plausible.

I don't think that you can deduct anything from one single instance, where *-ǵ and *-s are synonym forming endings; especially not even the fact that they are suffixes to a shorter root *le.

Another note: The Galician form lesen is a pure coincidence, the /s/ there does not come from the /g/ in leger, but from the ending, compare Catalan lleguessin with preserved /g/.

EDIT: The r in the Galician infinitve ler is also not part of the root, it goes back to the Latin infinitive ending -re. So no connection to German lernen "to learn" here.

P.S. For Some alternations in Protoindogermanic that occur with a greater frequency, see this question and its answers: Can these new etymological pairs of PIE roots be true?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.