search

funderburkjim / MWderivations

Derivations of headwords in the Monier-Williams (1899) dictionary
1 stars 1 forks source link

Using HBC headwords to resolve compounds #5

Open funderburkjim opened 8 years ago

funderburkjim commented 8 years ago

In the cpd1 method, an attempt is made to recognize a headword W whose key2 form is X-Y.

This requires that the first part, X, be the parent P of W.

The change made is that now X can be one the B or C variants of the parent P.

This change was made as a way to analyze uKA-saMBaraRa . I'll return to this example in another comment in this issue.

The net effect of this addition to the cpd1 logic is small (28 cases), but the details of derivation of many more cases are changed (960)

The case of avAk-ka (L=18508) illustrates some of the unexpected subtleties.

PREVIOUS analysis
2   18508   avAkka  avAk-ka m:f#A:n avAk-ka DONE    cpd4
CURRENT analysis
2   18508   avAkka  avAk-ka m:f#A:n avAk-ka DONE    cpd1

To understand the difference, here is the parentage of this avAkka word:

1   18500   avAYc   avAYc   m:f:n   
1C  18500.1  avAYcam  avAYc-am  
1B  18500.2  avAcI  avAcI  
1C  18500.3  avAk  avAk  
2  18501  avAk  avAk  
3  18501.1  avAkCruti  avAk-Cruti  
[7 more H3 cpds of avAk, L=18501]
2   18508   avAkka  avAk-ka m:f#A:n

The cpd1 method of the previous analysis looked at parent headword avAYc (L=18500), and failed. The cpd4 method doesn't worry about parents, so noticed that avAk was (some) headword (such as from L=18501, or 18500.3), counted 'ka' as a headword, and thereby got a derivation as cpd4.

In the current analysis, cpd1 considered any of headwords L=18500 through 18500.3 as candidate parent headwords, found avAk at L=18500.3, counted 'ka' as a headword, and succeeded.

Note1: the 'note' field is analysis2.txt shows 'cpd1:*' in the 960 cases where an 'alternate' parent headword was used, as with L=18505, avAkka example.

Note2: The ka in avAk-ka is a Whitney suffix, so what did the wsfx1 analysis fail? The answer is that wsfx1 uses only the 'old' definition of parent, as did the previous version of cpd1.

funderburkjim commented 8 years ago

Returning to the case of uKA-saMBaraRa.

This STILL is on the TODO list (although uKA is now a parent.

The reason is that there are inconsistencies in MW's use of anusvAra.

In particular, the headword spelling of saMBaraRa in MW does not use anusvAra: samBaraRa.

Not sure how to adjust for this in the derivations.

funderburkjim commented 8 years ago

Here's another similar M/m inconsistency:

pretasamkxpta but saMkxpta

gasyoun commented 3 years ago

This STILL is on the TODO list (although uKA is now a parent.

@funderburkjim still there, right?