compare to Huet, verbs, present conjugation passive

[funderburkjim]

This issue summarizes differences and similarities noticed in the comparison of verb conjugations; the methodology is described in the readme for huetcompare/verbs-prim-pas.

[funderburkjim]

root comparison

huet_stems_passive.txt summarizes the roots that appear in primary passive conjugations in the SL_roots.xml file from Huet.

pysan_stems_passive.txt summarizes the roots that appear in the dcpforms-MW-verb.txt of the pysan system. These were developed ca. 2004 from the Cologne digitization of the Monier-Williams as it then existed. A similar, more current, and in some ways more complete source of roots from MW can be found in this MWvlex file.

The two lists of roots are compared in compare_stems_passive.txt.

This comparison is organized into 3 categories:

• 520 cases of roots which appear in both sources
• 1094 cases of roots which appear only in the pysan source
• 37 cases of roots which appear only in the Huet source.

It would be interesting to delve further into these differences, and to make use of the probably more accurate MW data from the MWvlex source mentioned above. It is likely that most of the 37 Huet-only roots are represented in MW, but with slightely different spelling.

[funderburkjim]

conjugation table comparison

huet_conj_tables_passive.txt contains 2228 conjugation tables derived from Huet's SL_roots.xml.

For each root, there are 4 conjugation tables for the 4 present tenses, for which Huet uses theses abbreviations.

• pr = present tense
• ip = imperative
• op = optative
• im = imperfect

Each of these appears on one line of the file.

There are 557 roots (in huet_stems_passive.txt), and 4*557 = 2228, so the number of conjugation tables is as expected.

A comparison of the conjugation algorithms of Huet and pysan was made without taking into account the root differences of the two systems. Thus, for each of the Huet conjugation tables, a computation was done to derive a pysan conjugation, using the given root and tense as appearing in the Huet conjugation table entry. The result is pysan_conj_tables_passive.txt.

The part of the pysan system for computing conjugation tables for the passive take a conjugation class parameter; I arbitrarily set this to 1. While the generation of a passive stem generally is independent of the conjugation class of the root, I suspect that some of the logic in the pysan system may use the class of the root. This detail of the pysan system needs to be examined more closely.

Note that Huet does not compute passives for any of the denominative verbs.

compare_conj_tables_passive.txt summarizes the similarities and differences in the corresponding conjugations from the two files just mentioned. Some statistics resulting from the comparison:

• 1993 (89% of the 2228) passive conjugations are identical in the two systems.
  • The 'r-H' adjustment used in the present system conjugation comparisons was not required in this comparison of passive conjugations.
• 235 conjugations differ in one or more details.

[funderburkjim]

This is all the summary comments for now.

[gasyoun]

It is likely that most of the 37 Huet-only roots are represented in MW, but with slightely different spelling.

Yeah, guess nothing unique. Because Huet includes only real life roots that he has seen in text himself as far as I remember.

• The 'r-H' adjustment used in the present system conjugation comparisons was not required in this comparison of passive conjugations.

Would be strange if otherwise.