Part nineteen 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).

It’s now time to do for the middle forms what we did for the actives in part 16, namely come up with the rules to help disambiguate inflectional classes. These were sketched out in theory in part 14 but now it’s time to actually write the rules and test them in code against the SBLGNT.

This is what my Python script does:

INF:Xεσθαι or 3SG:Xεται or 2PL:Xεσθε is PM-1 if lemma ends in ω or ομαι
PM-7 if lemma ends ημι
1SG:Xομαι or 1PL:Xόμεθα or 3PL:Xονται is PM-8 if lemma ends in δίδομαι
PM-1 if lemma ends in ω or otherwise ends in ομαι
1SG:Xοῦμαι or 3PL:Xοῦνται is PM-2 if lemma ends in έω or έομαι
PM-3 if lemma ends in όω or όομαι
1SG:Xῶμαι or 1PL:Xώμεθα or 3PL:Xῶνται is PM-5 if lemma ends in χράομαι
PM-4 if lemma otherwise ends in άομαι
2SG:Xῇ is PM-2 if lemma ends in έω or έομαι
PM-5 if lemma ends in άομαι
1PL:Xύμεθα is PM-2 if lemma ends in έω or έομαι
PM-3 if lemma ends in όω or όομαι (not needed in SBLGNT)
PM-5 otherwise (not needed in SBLGNT)
3SG:Xεῖται or 2PL:Xεῖσθε is PM-2 if lemma ends in έω or έομαι
PM-11 if lemma ends in εῖμαι
1PL:Xείμεθα is PM-11 if lemma is κεῖμαι
PM-11-COMPOUND otherwise (not needed in SBLGNT)
INF:Xεῖσθαι is PM-2 if lemma ends in έω or έομαι
PM-11 if lemma is κεῖμαι (not needed in SBLGNT)
PM-11-COMPOUND otherwise
INF:Xῆσθαι is PM-10-COMPOUND if lemma is κάθημαι
PM-5 otherwise (not needed in SBLGNT)

I decided to cover a bunch of ambiguities not specifically needed by the SBLGNT—not strictly necessary but it will help when the script is extended to run on a larger corpus.

Note the special-casing of δίδομαι, κεῖμαι, κάθημαι, and χράομαι. χράομαι is an example, like ζάω in part 16, that is misleadingly lemmatized with an alpha. More on that later!

We now have an inflectional class for all 820 present middle infinitive or indicative forms in the MorphGNT SBLGNT.

You can download the entire output of my Python script here.

Are there multiple classes for a particular lexeme (like there was in the active)?

Two of the 167 lexemes show multiple classes:

  • δύναμαι: PM-9 normally but a 2SG:δύνῃ that comes up as a PM-1 (PM-9 would predict a Xασαι)
  • κάθημαι: PM-10-COMPOUND normally but a 2SG:κάθῃ that comes up as a PM-1 (PM-10-COMPOUND would predict a Xησαι)

If κάθῃ were καθῇ, we’d have the possibility of reanalysis as a PM-5 and it’s still possible that’s what’s going on and the accentuation just doesn’t reflect that.

δύνῃ for δύνασαι is somewhat less expected and it should be noted that both forms appear in the SBLGNT, sometimes within the same author. That the PM-4 2SG all show up with an un-contracted ᾶσαι adds slightly more mystery.

For now we’ll leave δύνῃ and κάθῃ as PM-1 but we revisit them later.

In the next part, we’ll look at counts for the present middles across the SBLGNT.