Finnish
- apertium-fin is a conversion from omorfi: omorfi, large coverage, experimental
- giella-fin is in giellatekno repository, more stable
Contents
Grammar stuff
Common formulas:
Numeral fix
People working on morpholocally poor language falsely assume that you should use
sg/pl
distinction in numerals with assigning the word "one" to singular always and any other words to plural. As we know though, the plural singular distinction is morphological and not lexical, and it can be used in all numbers, so we need to fix this in most non-uralic–Finnish pairs like so:
Negations
Finnish uses negation verb which needs to be translated from many languages from negation and verb together.
Possession structure
Finnish does not have idiomatic verb for possession, if apertium language to translate from has vbhaver it can be translated to omistaa initially and re-organised into adessive of owner and copula on.
Adposition to case suffix
Finnish uses semantic cases for what many e.g. IE languages use adpositions:
- houses -> talot
- on houses -> taloissa
- into houses -> taloihin
etc.
<section-def-macros>
<def-macro n="adp-mangler" npar="1">
<choose>
<!-- adp to case mappigngs -->
<!-- based on adp lexeme only -->
<when>
<test>
<equal><clip pos="1" side="sl" part="lem"/><lit v="I"/></equal>
</test>
<let>
<var n="adpcase"/>
<lit-tag v="ine"/>
</let>
<let>
<var n="maybeadp"/>
<lit v=""/>
</let>
</when>
<when>
<test>
<equal><clip pos="1" side="sl" part="lem"/><lit v="i"/></equal>
</test>
<let>
<var n="adpcase"/>
<lit-tag v="ine"/>
</let>
<let>
<var n="maybeadp"/>
<lit v=""/>
</let>
</when>
<when>
<test>
<equal><clip pos="1" side="sl" part="lem"/><lit v="fra"/></equal>
</test>
<let>
<var n="adpcase"/>
<lit-tag v="ela"/>
</let>
<let>
<var n="maybeadp"/>
<lit v=""/>
</let>
</when>
...
...
...
<rule comment="adp noun">
<pattern>
<pattern-item n="adp"/>
<pattern-item n="noun"/>
</pattern>
<action>
<call-macro n="adp-mangler">
<with-param pos="1"/>
</call-macro>
<out>
<chunk name="adpnoun" case="caseFirstWord">
<tags>
<tag><lit-tag v="NP"/></tag>
<tag><var n="adpcase"/></tag>
</tags>
<lu>
<clip pos="2" side="tl" part="lem"/>
<clip pos="2" side="tl" part="a_noun"/>
<clip pos="2" side="tl" part="a_number"/>
<var n="adpcase"/>
</lu>
<b pos="0"/>
<lu>
<var n="maybeadp"/>
</lu>
</chunk>
</out>
</action>
</rule>
The effective range of adposition to suffix mapping is a noun phrase:
- (a) green colourless dream -> vihreä väritön uni
- in (a) green colourless dream -> vihreässä värittömässä unessa
but proper noun phrase is appositive:
- for Donald Trump -> Donald Trumpille
- for president Trump -> presidentti Trumpille
in Nokia Finnish a reverse pattern is used (because computers cannot inflect variables a proxy word is inflected):
- user -> käyttäjä
- to user Joe -> käyttäjälle Joe
- file -> tiedosto
- into file thesis.doc -> tiedostoon thesis.doc
it would not be incorrect to use proper inflection in these cases.
