Dickinson College's digitization of the grammar text by Goodell seems to suggest that -τρον and -θρον are synonyms. We also have πτολίεθρον, where it looks to me like the suffix is -εθρον (unless this is somehow based on one of the plural forms of πτόλις that has an ε). Is the choice of suffix explainable based on some phonetic rule, possibly related to ease of articulation or historical evolution? I would have found ἄρτρον and πτολίθρον just as easy to pronounce as the actual forms.

It's cool to imagine that in some parallel universe, hydrogen atoms have electhrons in them.

  • 3
    Then probably elechthrons!
    – Cerberus
    Commented Oct 21, 2021 at 2:59

2 Answers 2


The endings -τρον, -θρον (and actually, a bunch of others also) are thought to share a common origin, but the origin of the τ/θ difference in particular is hypothesized to involve Proto-Indo-European laryngeals, which were lost in Greek, making it likely that there was no simple synchronic explanation by that point.

Based on the results of Greek Dictionary Headword Search Results at Perseus, words ending in -τρον are more common than words ending in -θρον in general, and -τρον is always used instead of -θρον after an obstruent consonant (the three clusters starting with obstruents that actually appear are -κτρον, -πτρον, -στρον; no words with the corresponding endings -φθρον, -χθρον, -σθρον appear to exist). I don't know whether an Ancient Greek speaker would have found this coincidental.

You can likely find other more recent sources regarding the situation in PIE/pre-PIE, but the one I have read is The Proto-Indo-European Instrument Noun Suffix *-tlom and its Variants, by Birgit Anette Olsen (1988), which suggests that the initial consonant in the -θρον type comes from metathesis of a stem-final laryngeal, forming a -tH- cluster that turned into a voiceless aspirate *tʰ. (More specifically, the laryngeals causing this change are specified to be *h1 and *h2).

Because the outcome of the derived voiceless aspirate *tʰ is the same in Greek and Latin as the outcome of the more common and basic voiced aspirate *dʰ, you'll also see the PIE suffix that developed to -θρον reconstructed with dʰ. Olsen describes dʰ as the "traditional" PIE reconstruction for that variant of the suffix, listing the following variant forms at the start of the paper:

*-tro-/*-tlo-/*-dʰro-/*-dʰlo- and *-trah2/*-tlah2/*-dʰrah2/*-dʰlah2

I'm not sure where the ε comes from in πτολίεθρον.

  • Olsen mentions πτολίεθρον in 7.3.7 as "quite obscure."
    – user3597
    Commented Oct 21, 2021 at 14:32

Asteroides has done an admirable job of answering the precise question as posed. This self-answer is just a follow-up based on the Olsen monograph to give a bit of the broader picture that I had not realized when I formulated my question, and to give a less technical presentation with some concrete examples in Greek.

In Greek, this family of PIE suffixes leads not just to τρον and θρον but also to τλον, θλον, θρη, and θλη. In Latin, it produces all kinds of things like trum, brum, bra, crum, culum, and clum.

The rules that Olsen proposes are I think, in dumbed-down form:

(1) If the root ends in s, then the ending has tr.

ἄγκιστρον, fish hook

This rule is absolute, and to me as an English speaker it seems to be closely related to ease of articulation. I find it very hard to say the consonant cluster στλ, and to a lesser extent σθρ and σθλ.

(2) Else, if the root has a liquid, then the ending has tr or θρ.

ἄροτρον, plow

πτολίεθρον, city

(3) Or if neither 1 or 2 applies, then we get tl or θλ.

χύτλον, libation, that which is poured

ἄεθλον, contest

γενέθλη, race, birthplace

There are many counterexamples to 3 in Greek:

σκῆπτρον, scepter

κέντρον, point

She explains this by saying that Greek developed a later preference for τρον, and that the vocabulary involved is often technical terminology, "later forms," or "foreign elements."

Your Answer

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