Talk:Separable verbs

From Apertium
Revision as of 11:00, 11 December 2009 by 79.156.80.145 (talk)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Example code[edit]

<?xml version="1.0" encoding="utf-8"?>
<transfer>

<section-def-cats>
  <def-cat n="default"> <!-- Catch all words -->
    <cat-item tags="*"/>
  </def-cat>
  <def-cat n="sepverb">
    <cat-item tags="verb.sep.*"/>
  </def-cat>
  <def-cat n="sepverb_part">
    <cat-item tags="part.sep"/>
  </def-cat>
</section-def-cats>
<section-def-attrs>
  <def-attr n="tags">
    <attr-item tags="vblex.pri.p2.pl"/>
    <attr-item tags="vblex.pri.p2.sg"/>
    <attr-item tags="vblex.pri.p1.pl"/>
    <attr-item tags="vblex.pri.p1.sg"/>
  </def-attr>
</section-def-attrs>
<section-def-vars>
  <def-var n="buffer" v=""/>
  <def-var n="verb_lemh"/>
  <def-var n="verb_lemq"/>
  <def-var n="verb_rest"/>
  <def-var n="verb_seen_sep"/>
  <def-var n="verb_seen_part"/>
</section-def-vars>

<section-rules>

  <rule comment="REGLA: SEP PART">
     <pattern>
       <pattern-item n="sepverb_part"/>
     </pattern>
     <action>
       <let>
         <var n="verb_seen_part"/>
         <lit v="true"/>
       </let>
       <let>
         <var n="verb_lemh"/>
         <concat>
           <var n="verb_lemh"/>
           <clip pos="1" side="sl" part="lem"/>
         </concat>
       </let>
     </action>
  </rule>

  <rule comment="REGLA: SEP VERB">
     <pattern>
       <pattern-item n="sepverb"/>
     </pattern>
     <action>
       <let>
         <var n="verb_seen_sep"/>
         <lit v="true"/>
       </let>
       <let>
         <var n="verb_lemh"/>
         <clip pos="1" side="sl" part="lemh"/>
       </let>
       <let>
         <var n="verb_lemq"/>
         <clip pos="1" side="sl" part="lemq"/>
       </let>
       <let>
         <var n="verb_rest"/>
         <clip pos="1" side="sl" part="tags"/>
       </let>
     </action>
  </rule>

  <rule comment="REGLA: DEFAULT">
     <pattern>
       <pattern-item n="default"/>
     </pattern>
     <action>
       <choose>
         <when>
           <test>
             <and>
               <equal>
                 <var n="verb_seen_sep"/>
                 <lit v="true"/>
               </equal>   
               <equal>
                 <var n="verb_seen_part"/>
                 <lit v="true"/>
               </equal>   
             </and>
           </test>
           <out>
             <lu>
               <var n="verb_lemh"/>
               <var n="verb_lemq"/>
               <var n="verb_rest"/>
             </lu>
             <b/>
             <lu>
               <var n="buffer"/>
             </lu>
           </out>
         </when>
         <when>
           <test>
             <and>
               <equal>
                 <var n="verb_seen_sep"/>
                 <lit v="true"/>
               </equal>   
               <not>
                 <equal>
                   <var n="verb_seen_part"/>
                   <lit v="true"/>
                 </equal>   
               </not>
             </and>
           </test>
           <let>
             <var n="buffer"/>
             <concat>
               <var n="buffer"/>
               <lit v="^"/>
               <clip pos="1" side="sl" part="whole"/>
               <lit v="$"/>
             </concat>
           </let>
         </when>
         <otherwise>
           <out>
             <lu>
               <clip pos="1" side="sl" part="whole"/>
             </lu>
           </out> 
         </otherwise>
       </choose>
     </action>
  </rule>

</section-rules>
</transfer>

Some discussion on this topic[edit]

Summary
  1. vblex (breek__vblex) "normal verbs" — these take the normal ge- prefix for forming the past participle, e.g. "gebreek", "geskryf", "gekyk", etc.
  2. vblex.uns (ontwikkel__vblex) "normal" verbs with unstressed prefixes, e.g. be-, ont-, her- — these do not take the ge- prefix, and have the same form in present, infinitive and past participle.
  3. separable figurative verbs
  4. separable verbs
Discussion
  • What is the deal with stuff that ends with ~d/~de like verbree/d — are these actually different participles?
(20:12:13) ftyers: btw, quick afrikaans question: "astronomers suppress the discovery" = "sterrekundiges druk die ontdekking onder." ?
(20:12:58) anon  : sterrekundiges onderdruk die ontdekking
(20:13:07) ftyers: hmm
(20:13:20) anon  : some seperable verbs are not always seperated
(20:13:22) ftyers: when would you use druk NP onder (if ever?)
(20:13:36) anon  : druk NP onder = sounds more literal to me
(20:13:41) anon  : onderdruk NP = more figurative
(20:13:50) ftyers: is the former valid ?
(20:13:55) anon  : the astronomers do not physically press down
(20:14:01) anon  : they press down figuratively
(20:14:02) ftyers: aye
(20:14:12) anon  : which former?
(20:14:23) ftyers: the druk NP onder combination
(20:14:31) ftyers: can it mean "suppress" or must it mean "press under" ?
(20:14:59) anon  : druk NP onder can't mean "suppress".
(20:15:08) ftyers: ok
(20:15:30) anon  : other examples:
(20:15:36) anon  : ondersoek = investigate
(20:15:41) ftyers: how about aansoek = apply ?
(20:15:45) anon  : soek onder = seek under(neath)
(20:16:12) anon  : no, but "aansoek" is not a verb... it requires aother verb, "doen".  To apply = om aansoek te doen
(20:16:25) ftyers: ah ok
(20:16:32) anon  : onderduik = to escape (the police)
(20:16:36) anon  : duik onder = dive under
(20:17:10) ftyers: can you give me some opposite examples ?
(20:17:27) anon  : I don't understand ?
(20:17:46) ftyers: some examples of where a separable verb must be split
(20:18:53) anon  : tricky....
(20:19:38) anon  : most seperable verbs combine with a preposition, and it depends mostly on the sentence construction whether it is seperate or joined.
(20:20:02) ftyers: hmm
(20:20:53) ftyers: does it depend on if the verb has another meaning or not ?
(20:20:56) ftyers: e.g. aankondig
(20:20:58) ftyers: = announce
(20:21:04) ftyers: but 'kondig' does not mean anything on its own
(20:21:19) anon  : could be...
(20:21:32) ftyers: let me think of another example
(20:21:37) anon  : opspoor, spoor op, but spoor does mean track
(20:21:47) ftyers: ok, aanrand
(20:21:51) ftyers: = attack
(20:21:54) ftyers: 'rand' doesn't mean anything
(20:22:03) ftyers: aanklag
(20:22:18) anon  : klag does mean something (klagte = complaint)
(20:22:27) ftyers: but not as a verb? 
(20:22:34) anon  : quite right
(20:23:46) anon  : yes but aanklag isn't a verb either
(20:23:54) ftyers: it isn't ?
(20:24:09) anon  : aankla = to report someone
(20:24:17) anon  : aanklag = complaint/report about someone
(20:24:21) ftyers: http://www.rieme.co.za/woordeboek.asp?l=1&woord=aanklag
(20:24:24) ftyers: i don't know if this is any good
(20:24:35) ftyers: but i understood it also meant "to charge someone" (with an offence)
(20:25:03) anon  : to charge someone = iemand aankla
(20:25:22) anon  : the URL you gave only gives nouns (some of them look like verbs, but they're nouns)
(20:25:36) anon  : charge = noun in this case
(20:25:36) ftyers: hmm quite right
(20:26:06) ftyers: i had it marked as 'check' aswell :)
(20:26:08) ftyers: glad i did
(20:27:43) anon  : one way to spot something like "ondersoek" is that the past tense would be "het ondersoek" or "het geondersoek", not "het onder gesoek".
(20:28:06) anon  : het ondersoek = investigated
(20:28:10) anon  : het onder gesoek = looked under
(20:28:31) ftyers: how about "aanrand" (aangerand)
(20:29:10) ftyers: e.g. "Polisie rand glo sakeman aan"
(20:29:39) anon  : yes... I'm afraid we might have missed each other here.
(20:29:52) ftyers: quite possible :)
(20:30:02) anon  : aanrand and ondersoek are two types of words -- ondersoek looks like a seperable verb but isn't.
(20:30:06) ftyers: aha
(20:30:12) anon  : aanrand is a seperable verb
(20:30:26) ftyers: and onderdruk is not a separable verb ?
(20:30:35) anon  : ondersoek looks like a seperable verb beacue it could be a seperatble verb if it had a different meaing.
(20:30:43) anon  : onderdruk = to suppress = not seperable
(20:30:47) ftyers: ok
(20:30:52) anon  : onderdruk = to press down = seperable
(20:31:16) anon  : so it is important to know the meaning
(20:31:19) ftyers: aye
(20:31:36) anon  : (although I must say that when I write "press down" I would probably write "onder druk")
(20:32:01) anon  : Hy wou haar kop onder druk (he wanted to press her head under)
(20:32:27) ftyers: and "he wanted to suppress her story" ?
(20:32:42) anon  : Hy wou haar storie onderdruk.
(20:33:04) ftyers: and "he suppresses her story" ?
(20:33:06) anon  : (this is where computers can't tell context -- a woman has a head, but not a story)
(20:33:17) ftyers: (sorry, should have given a simpler example)
(20:33:18) anon  : Hy onderdruk haar storie
(20:33:38) ftyers: ok
(20:33:56) ftyers: to be honest, i expected to have to go over verbs a few times before i understood properly
(20:34:23) ftyers: but is there an easy way to find out if a verb is "truely separable" or not ?
(20:34:25) anon  : unfortunately afrikaans is less systematic than esperanto
(20:34:30) ftyers: :)
(20:35:04) anon  : not really... unless you have a list of non-seperable verbs
(20:35:28) ftyers: can you think of a verb with onder- that is 'truely' separable ?
(20:36:03) anon  : what do you mean by "truly"?  Do you mean that it will always be seperatble?
(20:36:10) ftyers: yes
(20:36:11) ftyers: well
(20:36:17) anon  : difficult
(20:36:18) ftyers: like aankondig
(20:36:25) anon  : ondersteun = figurative
(20:36:31) anon  : ondergrawe = figurative
(20:36:38) ftyers: figurative isn't a verb o__O
(20:36:57) anon  : see, all the exapmles I can think of, are not only seperable
(20:37:09) ftyers: could you say for example: "sterrekundiges aankondig die ontdekking."
(20:37:15) anon  : no
(20:37:22) ftyers: ok
(20:37:32) anon  : but I'm trying to think of ones with "onder"
(20:37:59) ftyers: k
(20:39:55) anon  : I'm looking at a list of words here that start with "onder", and I still can't find any truly seperable verbs with onder
(20:40:12) ftyers: ok
(20:40:42) ftyers: how about onderloop ?
(20:40:48) anon  : there are a few verbs that have the form onder + ge + something
(20:40:51) ftyers: aha
(20:40:56) ftyers: thats what i was just looking at :)
(20:41:07) anon  : but you can't use them in the form something NP onder
(20:41:14) ftyers: ah
(20:41:32) anon  : onderwaardeer (undervalue)
(20:41:56) anon  : onderspoel (to erode from underneath)
(20:42:28) ftyers: ok
(20:42:30) anon  : aha... one exception:  onderkry
(20:42:35) ftyers: exceptions! :D
(20:42:38) ftyers: ok
(20:42:46) ftyers: kry = get
(20:42:48) anon  : but  onderkry has a figurative meaning only
(20:42:50) ftyers: onderkry = get under
(20:42:51) ftyers: ?
(20:43:06) anon  : onderkry = to beat, to lose against
(20:43:08) ftyers: ok
(20:43:29) anon  : jy moenie dat die probleme jou onderkry nie = you mustn't let the problems get you down
(20:43:55) anon  : die probleme kry hom onder = the problems are getting him down, the problems are causing him to fail
(20:44:50) ftyers: ok, so kinds of verbs i have come across so far:
(20:45:12) ftyers: 1. "truely separable" ones, e.g. aankondig
(20:45:21) ftyers: 2. "partially separable" ones, e.g. onderkry
(20:45:29) ftyers: 3. "separable looking" ones, e.g. ondersoek
(20:46:02) anon  : 3.1 seperable if literal meaning
(20:46:10) anon  : 3.2 not seperable if figurative meaning
(20:46:19) ftyers: 4. "non-separable", e.g. registreer
(20:47:02) ftyers: 1. take the past like het NP aan-ge-kondig
(20:47:13) ftyers: 2. take the past like het NP onder-ge-kry
(20:47:21) ftyers: 3. take the past like het NP ondersoek
(20:47:35) ftyers: 4. take the past like het NP geregistreer
(20:47:46) anon  : 3.1 sometimes het NP ge-ondersoek, depending on whether the author is in love with "ge".
(20:48:23) ftyers: ones in be- are almost always like 3.
(20:48:28) ftyers: only they can't take ge-
(20:48:42) ftyers: same goes for ones in ver-
(20:49:07) anon  : yes, it was an old spelling rule... be, ge, her, ver, ont
(20:49:23) ftyers: e.g. "i make that" = "ek vervaardig dat", "i made that" = "ek het dat vervaardig"
(20:49:38) anon  : yes
(20:49:44) anon  : except that "dat" is dutch
(20:49:47) ftyers: oops
(20:49:48) ftyers: sorry :)
(20:49:59) anon  : i make that = ek vervaardig dit
(20:50:01) ftyers: dit
(20:50:02) ftyers: ok
(20:50:11) anon  : i make it = ek vervaardig dit
(20:50:29) anon  : i make that thing = ek vervaardig daardie ding
(20:50:44) anon  : (can't use daardie without a noun)
(20:50:50) ftyers: k
(20:51:00) ftyers: ok, this makes things clearer
(20:51:11) ftyers: the ones that will cause the most trouble are 1 & 2
(20:51:15) anon  : strictly speaking you can't use "hierdie" without a noun either, but under influence of English it is more ocmmon now
(20:51:30) anon  : this house is big = hierdie huis is groot
(20:51:44) ftyers: but splitting verbs into the categories may prove difficult
(20:51:46) anon  : this is big = hierdie [ding] is groot
(20:51:54) anon  : yes
(20:52:45) anon  : I think most of the figurative #3 words woudl have seperate entries in a dictionary
(20:52:50) ftyers: yes
(20:52:58) ftyers: that won't really be a problem
(20:53:06) ftyers: but figurative #3 are like #2 words ?
(20:53:10) ftyers: *Verbs
(20:54:07) anon  : figurative #3 verbs can't be split
(20:54:12) ftyers: ah
(20:54:12) ftyers: sorry
(20:54:13) anon  : literal #3 can be split
(20:54:15) ftyers: i mean literal
(20:54:17) ftyers: yes
(20:54:18) ftyers: right
(20:55:16) anon  : do you have a long list of afrikaans words (without meanings)?
(20:55:34) ftyers: yes
(20:55:38) anon  : k
(20:55:45) ftyers: 159710 afrikaans-meester-utf8.txt
(20:55:59) anon  : wspel.co.za = 350 afr words
(20:56:03) anon  : 350 k
(20:56:03) ftyers: they are forms, not lemmata
(20:56:17) anon  : okay
(20:56:40) ftyers: (meaning mine
(20:56:44) ftyers: don't know about the other