Part thirty of a tour through Greek inflectional morphology to help get students thinking more systematically about the word forms they see (and maybe teach a bit of general linguistics along the way).

To complete the imperfect active indicatives, there are just a few more tweaks we need to make.

Firstly, we need to add the compound versions of IA-10 and IA-11.

Secondly, because the form ἦστε (for the 2PL of ἦν) appears in one of our test grammars, we need to add that to IA-10. More on that whole paradigm later.

Thirdly, let’s just rename IA-6 to IA-6a for consistency with how we named the present once we decided to include the υ.

That results in this update to IA-6a through IA-11-COMP.

  IA-6a IA-7 IA-8 IA-9 IA-9b IA-10 IA-10-COMP IA-11 IA-11-COMP
1SG Xῡν Xην/Xειν Xουν Xην Xην ἦ/ἦν Xῆ/Xῆν ᾖα/ᾔειν Xῇα/Xῄειν
2SG Xῡς Xεις Xους Xης Xης/Xησθα ἦς/ἦσθα Xῆς/Xῆσθα ᾔεις/ᾔεισθα Xῄεις/Xῄεισθα
3SG Xῡ Xει Xου ἦν Xῆν ᾔει/ᾔειν Xῄει/Xῄειν
1PL Xῠμεν Xεμεν Xομεν Xᾰμεν Xᾰμεν ἦμεν Xῆμεν ᾖμεν Xῇμεν
2PL Xῠτε Xετε Xοτε Xᾰτε Xᾰτε ἦτε/ἦστε Xῆτε/Xῆστε ᾖτε Xῇτε
3PL Xῠσᾰν Xεσᾰν Xοσᾰν Xᾰσᾰν Xᾰσᾰν ἦσᾰν Xῆσᾰν ᾖσᾰν/ᾔεσᾰν Xῇσᾰν/Xῄεσᾰν

And just for completeness, here’s the rest:

  IA-1 IA-2 IA-3 IA-4 IA-5
1SG Xον Xουν Xουν Xων Xων
2SG Xες Xεις Xους Xᾱς Xης
3SG Xε(ν) Xει Xου Xᾱ
1PL Xομεν Xοῦμεν Xοῦμεν Xῶμεν Xῶμεν
2PL Xετε Xεῖτε Xοῦτε Xᾶτε Xῆτε
3PL Xον Xουν Xουν Xων Xων

Does this handle all the forms in the MorphGNT plus our test grammars?


In Romans 3.13, we find ἐδολιοῦσαν, which does not match any of our patterns. What is happening here?

We have a contraction (suggesting IA-2 or IA-3 but, as indicated in the lemma, it’s an IA-3) but with a -σᾰν ending like we would expect in an athematic verb. Because the contraction would normally only happen with a theme-vowel, we don’t expect to see both -οῦ- and -σαν together.

If you look at IA-3 and IA-8 you can see they are indistinguishable in the singular. In fact IA-8 is acting like a thematic verb in the singular so there was already a merger happening between the classes. Further confusion about which endings to use in the plural makes sense, although here we have an interesting combination of distinguishers, combining the -οῦ- we might expect in an IA-3 plural with with the -σᾰν we expect in an athematic plural.

It’s worth pointing out that particular phenomenon is fairly common in the Septuagint and Romans 3.13 is a quote from the Septuagint. We can’t know for sure if Paul would normally have written ἐδολίουν instead but we can speculate that it’s like an American writer keeping British spelling in a quotation of a British author.

In our data, that’s the only form that fails to match. But there are others that exhibit a similar phenomenon that we should collect for completeness.

Twice in John 15 we find εἴχοσαν where we might expect εἶχον (and indeed do find outside of John). This is again common in the LXX.

More broadly (and not particularly characteristic of the LXX) is the replacement of athematic verbs with a thematic equivalent.

Twice in Acts we find ἐτίθουν where we would expect ἐτίθεσαν (acting like an IA-2 τιθέω).

Also in Acts we find ἀπεδίδουν (acting like an IA-3 διδόω) and both παρεδίδουν (Acts 27.1) and παρεδίδοσαν (Acts 16.4).

These athematic verbs are inflecting as if they were thematic. Note this actually causes a 1SG / 3PL ambiguity that wouldn’t otherwise exist.

There are other examples of athematic verbs inflecting as if thematic:

John 21.18 has ἐζώννυες with a theme vowel (becoming thematic IA-1 instead of IA-6a).

Matthew 21.8 has ἐστρώννυον with theme vowel (becoming thematic IA-1 instead of IA-6a).

Twice in Mark we find ἤφιε(ν) (becoming thematic IA-1 instead of IA-7 where we’d expect ἠφίει).

And, in different categories:

Acts 21.27 has uncontracted 3PL συνέχεον.

Acts 9.22 has 3SG συνέχυννεν as if the lemma were συγχύννω (and I’m tempted to, in fact, lemmatise it that way).

We also have cases of confusion between -αω and -εω verbs (which long happened in Greek dialects). ἠρώτουν in Matthew 15.23 looks like an IA-2 (or IA-3) even though ἠρώτα and ἠρώτων elsewhere suggest IA-4.

Unlike the usual confusions between inflectional classes we’ve seen above, though, there are no distinguisher patterns shared between IA-2 and IA-4 so the underlying cause is different.

A few other points to raise to round out the full set of imperfect active forms in our data:

In MorphGNT there is an open issue about ἔστηκεν in John 8.44. MorphGNT currently analyses it as an imperfect (it would be the imperfect of στήκω) but with the lemma ἵστημι (which would have a perfect of ἕστηκεν with rough breathing). This needs to be resolved in MorphGNT so doesn’t really effect our analysis of imperfect active forms here but I thought I’d mention it.

Another issue that should be considered in MorphGNT is εἰσῄει in Acts 21 should possibly be normalised with the movable nu as it’s a IA-11-COMP.

In the next post, we’ll go through resolving any remaining ambiguities in the imperfect active forms.