Difference between revisions of "User:Firespeaker/PSG transfer"
Jump to navigation
Jump to search
Firespeaker (talk | contribs) (Created page with "== An example == === two example sentences === телескоп жұмыс істейді = the telescope works = telescopio funciona адам сауда орталығын...") |
Firespeaker (talk | contribs) |
||
| (2 intermediate revisions by the same user not shown) | |||
| Line 32: | Line 32: | ||
</pre> |
</pre> |
||
* all I'm doing here is adding another verb to the lexicon: "жұмыс істе".v.ppv |
* all I'm doing here is adding another verb (with a new tag) to the lexicon: "жұмыс істе".v.ppv |
||
* ppv = verb taking PP-like argument (for Kazakh it's actually NP.loc) |
* ppv = verb taking PP-like argument (for Kazakh it's actually NP.loc) |
||
| Line 48: | Line 48: | ||
</pre> |
</pre> |
||
* you'd probably want some indication of finite VPs and V forms, optional Adj and Det in the NPs, and a whole bunch of other stuff... |
* you'd probably want some indication of finite VPs and V forms, optional Adj and Det in the NPs, agreement, and a whole bunch of other stuff... |
||
=== simple transfer === |
=== simple transfer === |
||
| Line 72: | Line 72: | ||
# note that agreement isn't implemented, but that could be done |
# note that agreement isn't implemented, but that could be done |
||
* The main thing here is that iv/ppv readings were disambiguated, meaning that your |
* The main thing here is that because things that didn't parse were discarded, iv/ppv readings were disambiguated, meaning that your bidix can contain the following mapping for use at the biltrans stage—and you wouldn't have to worry about lexical selection (and a lot of disambiguation has been done already too): |
||
** "жұмыс істе".v.iv = "funcionar".v.iv |
** "жұмыс істе".v.iv = "funcionar".v.iv |
||
** "жұмыс істе".v.ppv = "trabajar".v.ppv |
** "жұмыс істе".v.ppv = "trabajar".v.ppv |
||
Latest revision as of 09:29, 19 November 2014
Contents
An example[edit]
two example sentences[edit]
телескоп жұмыс істейді = the telescope works = telescopio funciona
адам сауда орталығында жұмыс істейді = the man works at the mall = el hombre trabaja en el centro comercial
current morph[edit]
Actual output of apertium-kaz (with newlines added for legibility):
^телескоп/телескоп<n><nom>/телескоп<n><attr>/телескоп<n><nom>+е<cop><aor><p3><pl>/телескоп<n><nom>+е<cop><aor><p3><sg>$ ^жұмыс істейді/жұмыс істе<v><iv><aor><p3><pl>/жұмыс істе<v><iv><aor><p3><sg>$^./.<sent>$
^адам/адам<n><nom>/адам<n><attr>/адам<n><nom>+е<cop><aor><p3><pl>/адам<n><nom>+е<cop><aor><p3><sg>$ ^сауда орталығында/сауда орталық<n><loc>/сауда орталық<n><px3sp><loc>/сауда орталық<n><loc>+е<cop><aor><p3><pl>/сауда орталық<n><loc>+е<cop><aor><p3><sg>/сауда орталық<n><px3sp><loc>+е<cop><aor><p3><pl>/сауда орталық<n><px3sp><loc>+е<cop><aor><p3><sg>$ ^жұмыс істейді/жұмыс істе<v><iv><aor><p3><pl>/жұмыс істе<v><iv><aor><p3><sg>$^./.<sent>$
proposed morph[edit]
^телескоп/телескоп<n><nom>/телескоп<n><attr>/телескоп<n><nom>+е<cop><aor><p3><pl>/телескоп<n><nom>+е<cop><aor><p3><sg>$ ^жұмыс істейді/жұмыс істе<v><iv><aor><p3><pl>/жұмыс істе<v><iv><aor><p3><sg>/жұмыс істе<v><ppv><aor><p3><pl>/жұмыс істе<v><ppv><aor><p3><sg>$^./.<sent>$ ^адам/адам<n><nom>/адам<n><attr>/адам<n><nom>+е<cop><aor><p3><pl>/адам<n><nom>+е<cop><aor><p3><sg>$ ^сауда орталығында/сауда орталық<n><loc>/сауда орталық<n><px3sp><loc>/сауда орталық<n><loc>+е<cop><aor>* <p3><pl>/сауда орталық<n><loc>+е<cop><aor><p3><sg>/сауда орталық<n><px3sp><loc>+е<cop><aor><p3><pl>/сауда орталық<n><px3sp><loc>+е<cop><aor><p3><sg>$ ^жұмыс істейді/жұмыс істе<v><iv><aor><p3><pl>/жұмыс істе<v><iv><aor><p3><sg>/жұмыс істе<v><ppv><aor><p3><pl>/жұмыс істе<v><ppv><aor><p3><sg>$^./.<sent>$
- all I'm doing here is adding another verb (with a new tag) to the lexicon: "жұмыс істе".v.ppv
- ppv = verb taking PP-like argument (for Kazakh it's actually NP.loc)
simple grammar[edit]
N: n
V: {v.iv, v.ppv}
NP.nom → N.nom
NP.loc → N.loc
VP → V.iv
VP → NP.loc V.ppv
S → NP.nom VP
- you'd probably want some indication of finite VPs and V forms, optional Adj and Det in the NPs, agreement, and a whole bunch of other stuff...
simple transfer[edit]
I forget how this works exactly, but you'd need to take care of the following at this stage:
- fix the order of the VP (easy)
- translate the NP.loc into a PP (maybe not too hard?)
- somehow add some determiners (??)
output disambiguated sentences[edit]
^телескоп/телескоп<n><nom>$ ^жұмыс істейді/жұмыс істе<v><iv><aor><p3><pl>/жұмыс істе<v><iv><aor><p3><sg>$^./.<sent>$
^адам/адам<n><nom>$ ^сауда орталығында/сауда орталық<n><loc>/сауда орталық<n><px3sp><loc>$ ^жұмыс істейді/жұмыс істе<v><ppv><aor><p3><pl>/жұмыс істе<v><ppv><aor><p3><sg>$^./.<sent>$
- note that agreement isn't implemented, but that could be done
- The main thing here is that because things that didn't parse were discarded, iv/ppv readings were disambiguated, meaning that your bidix can contain the following mapping for use at the biltrans stage—and you wouldn't have to worry about lexical selection (and a lot of disambiguation has been done already too):
- "жұмыс істе".v.iv = "funcionar".v.iv
- "жұмыс істе".v.ppv = "trabajar".v.ppv
