[Attempto] Categories in the ACE lexicon

[Jörg Preisendörfer]

Hej Kaarel,

I see that sections 12.9 to 12.11 of the ACE Syntax Report 6.7 also 
offer some honey regarding that question. The boxes* there allude to how 
the morphological forms are disambiguated into lexical units.

*) [I don't call the boxes 'rules', because some boxes in the Syntax 
Report contain more than one production, and my guesswork is that when 
several productions occure in a single box of the Syntax Report, the box 
is essentially to be interpreted as a disjunction of the productions 
contained.]

Since I'm in the process of implementing ACE in another programming 
language's sociotope and am seeking to come up with a reasonable data 
model, I wonder what the name of a common superclass of the ACE lexical 
classes as layed out in the ACE Lexicon Specification could be, such 
that neither linguists nor software engineers are tempted to shed tears?

Since 'finsg' and 'infpl' each represent equal morphological forms of 
distinct lexical units, I assume that 'lexical unit' is not an entirely 
matching descriptive name for such a superclass.

What do you think?

Bests

-- J.



Am 13.03.22 um 12:27 schrieb Kaarel Kaljurand:
> Hi,
> 
> the spec also contains "formulas" like:
> 
>      tv_finsg(ThirdSgForm, LogicalSymbol).
>      tv_infpl(InfForm, LogicalSymbol).
>      tv_pp(PastPartForm, LogicalSymbol).
> 
> linking e.g. "finsg" to "ThirdSgForm" etc. I agree that it would be
> more clear to avoid such duplicate names (they are not used in the
> source code, see
> https://github.com/Attempto/APE/tree/master/prolog/lexicon), and be
> explicit about the linguistic theory, which is something like: (in
> ACE) the syntactic categories of "infinitive" and "plural" use the
> same morphological form, e.g. the sentences "John wants to run." and
> "At least 2 men run." both use the form "run".
> I don't remember the decisions on the naming conventions, and some of
> it pre-dates my involvement (i.e. pre 2004). The original lexicon was
> based on the work:
> 
> Alexandra Bünzli. AceLex — Lexikon für Ace, 2004. Diploma Thesis.
> Faculty of Arts, University of Zurich.
> http://attempto.ifi.uzh.ch/site/pubs/papers/lizarbeit_ABuenzli.pdf
> 
> But I think it was simplified later and names somewhat changed.
> 
> Best,
> Kaarel
> 
> On Thu, Mar 10, 2022 at 12:00 PM Jörg Preisendörfer
> <hallo at joerg-preisendoerfer.de> wrote:
>>
>>
>> Dear Tobias,
>>
>> thx for the prompt reply!
>>
>> The ACE lexicon specification actually doesn't say anything about the
>> rational of the mnemonics of 'finsg' and 'infpl'. In particular, 'finsg'
>> is pretty much the only mnemocode in the lexcicon specification that
>> doesn't have its roots addressed directly or alluded to in the text of
>> the specification.
>>
>> Of course, I can come up with a guess about how 'fin' and 'inf' differ,
>> but I think it would be better to be clear about the circumscription of
>> the two classes.
>>
>> To see why I'm interested in finding out about the intentions of those
>> mnemonics:
>>
>> The lexicon specification quietly assumes a certain theory of lexical
>> classes which is not made explicit (e.g. by naming a certain theory), at
>> least not in the specification document. This leads to the specification
>> not being self-contained.
>>
>> Presumably the specific theory employed for ACE holds no surprises, but
>> I'd still prefer someone involved saying something like "25 years ago,
>> we arrived at that mnemonics because of such and such reasons" rather
>> than guessing myself.
>>
>> Pls kindly bear with me in being picky about such aspects of the
>> documentation. :-) This is actually a relative simple case; there are
>> more ambiguities in the ACE Syntax Report which I will try to sort out next.
>>
>> Thx again,
>>
>> -- J.
>>
>>
>>
>>
>>
>>
>> Am 09.03.22 um 06:53 schrieb Tobias Kuhn:
>>> Dear Jörg,
>>>
>>> The documentation page of the lexicon format can be found here:
>>> http://attempto.ifi.uzh.ch/site/docs/ace_lexicon.html
>>>
>>> I hope that helps.
>>>
>>> Regards,
>>> Tobias
>>>
>>>
>>> On 08.03.22 18:08, Jörg Preisendörfer wrote:
>>>>
>>>>
>>>> Happy Women's Day everybody,
>>>>
>>>> in wrapping my head around ACE, some questions occured to me which I'd
>>>> like to sort out step by step.
>>>>
>>>> I am aware that the answers in many cases will be entirely obvious for
>>>> people who were involved in the development of ACE or even for every
>>>> linguist of some sort, but I'd nevertheless prefer your answers over
>>>> uninformed or informed guessing. You may want take that approach as an
>>>> expression of respect. :-)
>>>>
>>>>
>>>> To start with a simple one, here's a question refering to the word
>>>> catergories in the ACE lexicon,
>>>>
>>>> *
>>>> https://github.com/Attempto/APE/blob/master/prolog/lexicon/clex_lexicon.pl
>>>>
>>>>
>>>>
>>>> The lexicon uses the abbreviations 'finsg' and 'infpl' to mark verb
>>>> categories.
>>>>
>>>> Am I assuming right that 'finsg' stands for 'finite' and 'singular'
>>>> whereas 'infpl' stands for 'infinite' and 'plural'?
>>>>
>>>>
>>>> Thx
>>>>
>>>> -- J.
>>>> _______________________________________________
>>>> attempto mailing list
>>>> attempto at lists.ifi.uzh.ch
>>>> https://lists.ifi.uzh.ch/listinfo/attempto
>>> _______________________________________________
>>> attempto mailing list
>>> attempto at lists.ifi.uzh.ch
>>> https://lists.ifi.uzh.ch/listinfo/attempto
>> _______________________________________________
>> attempto mailing list
>> attempto at lists.ifi.uzh.ch
>> https://lists.ifi.uzh.ch/listinfo/attempto