The word Talo in Finnish means house. According to the wiktionary, it might be etymologically related to talas (boat-shelter). I was wondering if the word might have a common etymology with Greek Thalassa ("sea") or Thalamos ("bedroom").

I know it might be a long shot, and I know that there is not a known common ancestor of Proto Uralic and Proto Indo-European, but there is a small set of convergent vocabulary (for instance, in the personal pronouns), and I was wondering if this could be the case with this word. Could Talo and Thalamus actually be cognates?

  • 1
    did you randomly pick these words and if not what made you wonder if they might be related?
    – Ajagar
    Commented Sep 3, 2019 at 17:55
  • I assume it is phonetic resemblance. Taking into account that it is proven that we are all homo sapiens who originated in one continent and proven to have spread across the world and into space the scientific community assumes that words with similar phonemes in different languages have no etymological connection. That is like saying Swedes are in no way related to Koreaans.
    – Ajagar
    Commented Sep 3, 2019 at 17:59
  • Only because they live at great distance from each other. I have family in the middle east, Canada, Netherlands and many places I am not aware of going back in my family tree. We should look at language from different perspectives. It does not care for borders or nationalities. If a word was first written down in Greek did it exist in Greek first? It is a very important assumption and mistake to say Yes. Some words may definately be Greek in origin but Greek words had their origin too.
    – Ajagar
    Commented Sep 3, 2019 at 18:03
  • And people are in opposition to any explanation that is outside their own wrong scientific point of view. Work together, I keep saying that. Good question! +1 from me!
    – Ajagar
    Commented Sep 3, 2019 at 18:06

1 Answer 1


The problem is that both Greek words are probably not of Indogermanic origin. The case of θάλασσα is pretty clear-cut, the -σσ- cannot be inherited directly from Proto-Indogermanic and must be inherited from a pre-Greek substrate. The case of θάλαμος is less clear, but again a substrate origin is suggested.

The nature of the pre-Greek substrate remains unknown, but a connection to Hurro-Urartian has been proposed.

On the Finnish side it seems that talo lacks cognates in the Uralic languages (the Sami cognates are classified as loans from Finnish), and it is not an obvious loan word. This makes it difficult to track it back in time.

tl;dr With our current knowledge, the null hypothesis (the word forms are just chance coincidences) cannot be disproved.

  • See also linguistics.stackexchange.com/questions/32076/… on θάλασσα Commented Aug 27, 2019 at 15:25
  • Chance coincidence? Just say you don’t know where or how to look. It can be disproven but not with conventional methods.
    – Ajagar
    Commented Sep 2, 2019 at 21:17
  • 3
    @Ajagar The standard in science (linguistics as well as other disciplines) is to assume the null hypothesis until you have significant evidence to the contrary. Statistically, if you take two random words from random languages in the world, the probability that those two words are etymologically related is effectively zero. So it makes sense to assume that words are unrelated until you have enough evidence saying they are related, rather than the other way around.
    – Draconis
    Commented Sep 2, 2019 at 21:38
  • Well in this case the words chosen in the question are not random.
    – Ajagar
    Commented Sep 3, 2019 at 17:52
  • 1
    @Ajagar That's how it works, though: since the null hypothesis holds in 99.9999% of all cases, you have to assume it's true in 100%, then disprove it in 0.00001%. Which is significantly easier than the alternative: assume a connection in 100% of cases, then disprove that connection in 99.9999%. That's why the null hypothesis is such a useful scientific tool: it narrows down the field.
    – Draconis
    Commented Sep 3, 2019 at 19:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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