[Lingtyp] What do glossing labels stand for?

Sebastian Nordhoff sebastian.nordhoff at glottotopia.de
Wed Jan 27 12:30:44 UTC 2016


On 01/27/2016 12:58 PM, Volker Gast wrote:
> 
> With respect to glossing in general, I think that 'Advanced glossing' as
> proposed by HH Lieb and S Drude for the DOBES-programme makes a lot of
> sense (though the following description contains some unnecessary
> polemics):
> 
> http://dobes.mpi.nl/documents/Advanced-Glossing1.pdf
> 
> Lieb and Drude try to get away from the item-and-arrangement ideology
> that still underlies glossing conventions in typology. Category labels
> are clearly language-specific. (But then, the format targets language
> documentation, not typological studies.)
> 
> I would also regard the Leipzig Glossing Rules as a useful standard for
> abbreviations and the 'orthography' of glosses. And I agree that most
> glosses used in typological studies are no more than "reading aids", as
> Johanna seemed to imply. If they were to have any theoretical
> significance, or if they were to be interpretable in some way, we should
> at least indicate what combines with what within a word, and what
> relates to what (or what takes scope over what). For instance, in
> 
> adam-lar-a
> man-PL-DAT
> 
> there seems to be a hierarchical structure implied:
> [adam-lar]-a
> 
> '-lar' indicates a property of (the denotation of) 'adam', and '-a'
> indicates a property of (the denotation of) 'adam-lar'. There's probably
> no constituent '-lar-a' in that word, but that's not recoverable from
> the gloss.

in grammatical descriptions, the examples are typically embedded in a
larger discussion, which they exemplify and illustrate. Depending on
what the focus of that discussion is, linguistic information may be
backgrounded or foregrounded. One could add brackets for constituency
structure, or one could add a line with phonological features to
adam-lar-a, depending on whether this example is found in the phonology
chapter or the morphology chapter. One could also remove the hyphens if
the internal structure of the word does not add to the argument.

Jeff Good distinguishes between an example and an exemplar. An exemplar
is a cherry-picked example, which illustrates a particular point
particularly well. Exemplars then can have all kinds of didactically
useful extra material (Brackets, asterisks, indices, traces etc)

IGT in grammatical descriptions is typically of the exemplar-type, and
is contextualised by the surrounding prose. This also allows to recover
information not directly retrievable from the example. This is different
from (laguage documentation) corpora, where there is no tight connection
between the IMT example and any particular didactic point.

The Leipzig glossing rules historically targetted the exemplar-type.
They allow for conflation of categories, underspecification of
segmentation etc. Advanced Glossing is very different from that in that
its aim is to be a general purpose schema.

Best
Sebastian

PS: On a more technical note, I might add that stand-off annotation
seems to be preferable to inline annotation for general purpose schemas.
See https://github.com/xigt/xigt for stand-off annotation being applied
to IGT





> 
> Advanced glossing provides for word-internal structure and relations
> between the constituents of a word (though in a somewhat inelegant way).
> 
> Volker
> 
> Am 27.01.2016 um 11:09 schrieb Hedvig Skirgård:
>> Thanks Sebastian, I always appreciate those papers :).
>>
>> Just to be even clearer from my part as well, by a database I simply
>> here mean a non-linear source of information with multiple connections
>> possible. A wiki for example would qualify.
>>
>> /Hedvig
>>
>> *Hedvig Skirgård*
>> PhD Candidate
>> The Wellsprings of Linguistic Diversity
>>
>> ARC Centre of Excellence for the Dynamics of Language
>>
>> School of Culture, History and Language
>> College of Asia and the Pacific
>>
>> Rm 4203, H.C. Coombs Building (#9)
>> The Australian National University
>>
>> Acton ACT 2601
>>
>> Australia
>>
>> Co-char of Public Relations
>>
>> International Olympiad of Linguistics
>>
>> www.ioling.org <http://www.ioling.org>
>>
>>
>> On 27 January 2016 at 18:54, Sebastian Nordhoff
>> <sebastian.nordhoff at glottotopia.de
>> <mailto:sebastian.nordhoff at glottotopia.de>> wrote:
>>
>>
>>
>>     On 01/27/2016 02:18 AM, Stef Spronck wrote:
>>     > Hi Hedvig,
>>     >
>>     >
>>     >
>>     > Just to respond to your last point about preferring databases
>>     over grammar books: as someone originally trained as a typologist
>>     and then writing a fieldwork-based thesis, I agree that that
>>     experience changes the way in which you read grammars entirely. I
>>     also agree that there can never be enough collaboration between
>>     fieldworkers and typologists.
>>     >
>>     >
>>     >
>>     > But I don't think we should conflate the rightly increased focus
>>     on the accountability of grammar writers, resulting in the
>>     professionalisation of archiving practices, with the goals of
>>     grammatical description. The maturation of language documentation
>>     as a discipline, separate from language description, following
>>     Himmelmann's work has been an extremely important development. But
>>     linked corpora, as the product of language documentation, are not
>>     grammars. I think that a 'traditonal book grammar' as an
>>     intermediary between data repositories and typology has value,
>>     exactly because it explicitly serves to interpret the labels in
>>     our glosses and tries to account for a language as a system. This
>>     does introduce a distinction between the interpretation of glosses
>>     in desciptive grammars and in typology, as the many interesting
>>     contributions to the present discussion aim to address.
>>
>>     +1 Stef
>>
>>     It is important to make the distinction between a grammar as a
>> "set of
>>     rules" and a grammar as a didactic text genre (the grammatical
>>     description).  These have different communicative goals and different
>>     audiences.
>>     The "set of rules" part can be modeled as a typological database. The
>>     didactic part is normally not modeled, but running prose text.
>> Further
>>     information about this distinction and about ways to represent both
>>     types in computers can be found in the following publications
>>
>>     Nordhoff, Sebastian. 2008 Electronic Reference Grammars for Typology:
>>     Challenges and Solutions. JLDC 2(2). 296-324
>>    
>> http://scholarspace.manoa.hawaii.edu/bitstream/handle/10125/4352/nordhoff.pdf?sequence=7
>>
>>
>>     Nordhoff, Sebastian (ed.). 2012. Electronic Grammaticography (LD&C
>>     Special Publication No. 4). Manoa: University of Hawai‘i Press.
>>     http://nflrc.hawaii.edu/ldc/?p=263
>>
>>     Nordhoff, Sebastian and Harald Hammarström (2014).
>>     Archiving grammatical descriptions. In David Nathan & Peter K. Austin
>>     (eds) Language Documentation and Description, vol 12: Special Issue
>>     on Language Documentation and Archiving. London: SOAS. pp. 164-
>>     186 http://www.elpublishing.org/PID/143
>>
>>     Best
>>     Sebastian
>>
>>
>>
>>
>>     >
>>     >
>>     >
>>     > Best,
>>     > Stef
>>     >
>>     >
>>     >
>>     >
>>     >
>>     > Stef Spronck
>>     >
>>     > KU Leuven, Linguistics, research unit
>>     FunC<http://www.arts.kuleuven.be/ling/func> | KU Leuven
>>     profile<http://http//www.kuleuven.be/wieiswie/nl/person/00098925>
>>     | Personal website<http://people.anu.edu.au/stef.spronck/>
>>     >
>>     > ________________________________
>>     > Van: Lingtyp [lingtyp-bounces at listserv.linguistlist.org
>>     <mailto:lingtyp-bounces at listserv.linguistlist.org>] namens Hedvig
>>     Skirgård [hedvig.skirgard at gmail.com
>>     <mailto:hedvig.skirgard at gmail.com>]
>>     > Verzonden: woensdag 27 januari 2016 1:31
>>     > Aan: Östen Dahl
>>     > CC: lingtyp at listserv.linguistlist.org
>>     <mailto:lingtyp at listserv.linguistlist.org>
>>     > Onderwerp: Re: [Lingtyp] What do glossing labels stand for?
>>     >
>>     > Thank you everyone, in particular Nordhoff for that very clear
>>     list :D. It's exactly what I was thinking too.
>>     >
>>     > The Leipzig glossing rules do state that they are not absolute
>>     rules, but merely summing up already existing conventions. I.e.
>>     you can't really "apply the Leipzig glossing rules" without giving
>>     any more information, but just use it as a help in formulating
>>     your own glossing rules. Right? Or am I misreading it?
>>     >
>>     > I agree with Dahl though, a paragraph on comparative concepts vs
>>     lg-spec descriptive (or whatever terms you want to use) would be
>>     useful.
>>     >
>>     > Intertwined with this is that lg-spec authors of grammars often
>>     envisage certain readers, like typologists and sometimes look for
>>     "guidance" in typological literature - an enterprise that is
>>     problematic like we've discussed. Furthermore, there's the
>>     potential issues with expecting PhD students to write a grammar
>>     (sketch) in four years and with a quite restrictive page count.
>>     >
>>     > I've worked in a grammatical survey the last few years on
>>     african languages, and interacted a lot with fieldworkers working
>>     on those languages (extremely rewarding!!) and discussed their
>>     manuscripts with them. (This is something I highly-highly
>>     encourage every grammar-reading typologist to do if they aren't
>>     already). From my relatively limited experience reading grammars
>>     so far (compared to say Dahl or Dryer) there are some things that
>>     I've found more helpful as a reader and tried to suggest to the
>>     writers. Basically I'd like them to be longer, more repetitive,
>>     more examples and explicit in more assumptions. Preferably, for
>>     me, I'd like a grammar not to be book really, but rather a
>>     database with linked corpora. In connection to that, I'd like for
>>     there to be better standards in publishing corpora and getting
>>     proper credit for it.
>>     >
>>     > Might I also recommend this issue of LD&C that brings up new
>>     methods in grammar writing?
>>     > http://nflrc.hawaii.edu/ldc/?p=263
>>     >
>>     > /Hedvig
>>     >
>>     >
>>     > Hedvig Skirgård
>>     > PhD Candidate
>>     > The Wellsprings of Linguistic Diversity
>>     >
>>     > ARC Centre of Excellence for the Dynamics of Language
>>     >
>>     > School of Culture, History and Language
>>     > College of Asia and the Pacific
>>     >
>>     > Rm 4203, H.C. Coombs Building (#9)
>>     > The Australian National University
>>     >
>>     > Acton ACT 2601
>>     >
>>     > Australia
>>     >
>>     > Co-char of Public Relations
>>     >
>>     > International Olympiad of Linguistics
>>     >
>>     > www.ioling.org <http://www.ioling.org><http://www.ioling.org>
>>     >
>>     > On 27 January 2016 at 09:14, Östen Dahl <oesten at ling.su.se
>>     <mailto:oesten at ling.su.se><mailto:oesten at ling.su.se
>>     <mailto:oesten at ling.su.se>>> wrote:
>>     > OK, do I understand this correctly? The labels stand for
>>     language-specific categories, but normally we arbitrarily choose
>>     labels that are names of comparative concepts, without asserting
>>     "any relation between the morpheme being glossed and a comparative
>>     concept however defined (beyond the mnemonic usefulness).". But at
>>     the same time, according to the document at
>>     https://www.eva.mpg.de/lingua/resources/glossing-rules.php, the
>>     glosses are intended to "give information about the meanings and
>>     grammatical properties of individual words and parts of words".
>>     Can you do that without asserting any relation between the
>>     comparative concept identified by the label and the meaning of the
>>     item being glossed? The document says: "In many cases, either a
>>     category label or a word from the metalanguage is acceptable".
>>     Does this mean that lexical glosses are also only mnemonic?
>>     >
>>     > There is also a pedagogical problem here. There is no mention in
>>     the document of the distinction between descriptive categories and
>>     comparative concepts. The question is if people who write
>>     typological papers as well as those who read them understand the
>>     significance of glosses. I think there is a general tendency
>>     towards fundamentalism in most of us in the sense that we tend to
>>     take things more literally than they were intended to. So I
>>     suspect that most people who see the gloss DAT will think that it
>>     means that the author really thinks that the form in question is a
>>     dative, or at least matches some idea of what datives are like. Or
>>     that if the German word "Pferd" is glossed as 'horse', that means
>>     that it means 'horse'. In other words, it might be worth having
>>     some discussion in the document about these problems.
>>     >
>>     > östen
>>     >
>>     >
>>     > -----Ursprungligt meddelande-----
>>     > Från: Lingtyp [mailto:lingtyp-bounces at listserv.linguistlist.org
>>    
>> <mailto:lingtyp-bounces at listserv.linguistlist.org><mailto:lingtyp-bounces at listserv.linguistlist.org
>>
>>     <mailto:lingtyp-bounces at listserv.linguistlist.org>>] För William
>> Croft
>>     > Skickat: den 26 januari 2016 17:23
>>     > Till: lingtyp at listserv.linguistlist.org
>>    
>> <mailto:lingtyp at listserv.linguistlist.org><mailto:lingtyp at listserv.linguistlist.org
>>
>>     <mailto:lingtyp at listserv.linguistlist.org>>
>>     > Ämne: Re: [Lingtyp] What do glossing labels stand for?
>>     >
>>     > Exactly.
>>     >
>>     > Bill
>>     >
>>     > On Jan 26, 2016, at 1:11 AM, Sebastian Nordhoff
>> <sebastian.nordhoff at glottotopia.de
>>    
>> <mailto:sebastian.nordhoff at glottotopia.de><mailto:sebastian.nordhoff at glottotopia.de
>>
>>     <mailto:sebastian.nordhoff at glottotopia.de>>> wrote:
>>     >
>>     >> Dear list,
>>     >> - a language-specific category is a concept with a label chosen
>>     by the
>>     >> linguist.
>>     >> - the label is in principle arbitrary.
>>     >> - for mnemonic reasons, a label evocative of the concept being
>>     >> described is normally used.
>>     >> - since some labels are rather long, it is convenient to
>>     abbreviate them.
>>     >> - some abbreviations have several plausible expansions
>>     (SUPerlative,
>>     >> SUPeressive, SUPine)
>>     >> - a standardization of the match abbreviation-long label is
>>     therefore
>>     >> useful for disambiguation purposes. This is what the Leipzig
>>     glossing
>>     >> rules do in my opinion
>>     >> - the Leipzig glossing rules therefore match abbreviations with
>>     common
>>     >> concept labels. An author using a Leipzig gloss does, however, not
>>     >> assert any relation between the morpheme being glossed and a
>>     >> comparative concept however defined (beyond the mnemonic
>>     usefulness).
>>     >>
>>     >> Best wishes
>>     >> Sebastian
>>     >>
>>     >>
>>     >> On 01/25/2016 09:27 PM, Östen Dahl wrote:
>>     >>> Here is a question that I would like to pose to the members of
>>     the ALT list. If we accept the distinction between "descriptive
>>     categories" and "comparative concepts", what do the labels we use
>>     in glossing example sentences stand for - in particular, the
>>     labels defined in the Leipzig glossing rules? I have some thoughts
>>     about this myself but would like to hear what others think first.
>>     >>> östen
>>     >>>
>>     >>>
>>     >>> _______________________________________________
>>     >>> Lingtyp mailing list
>>     >>> Lingtyp at listserv.linguistlist.org
>>    
>> <mailto:Lingtyp at listserv.linguistlist.org><mailto:Lingtyp at listserv.linguistlist.org
>>
>>     <mailto:Lingtyp at listserv.linguistlist.org>>
>>     >>>http://listserv.linguistlist.org/mailman/listinfo/lingtyp
>>     >>>
>>     >> _______________________________________________
>>     >> Lingtyp mailing list
>>     >> Lingtyp at listserv.linguistlist.org
>>    
>> <mailto:Lingtyp at listserv.linguistlist.org><mailto:Lingtyp at listserv.linguistlist.org
>>
>>     <mailto:Lingtyp at listserv.linguistlist.org>>
>>     >>http://listserv.linguistlist.org/mailman/listinfo/lingtyp
>>     >
>>     > _______________________________________________
>>     > Lingtyp mailing list
>>     > Lingtyp at listserv.linguistlist.org
>>    
>> <mailto:Lingtyp at listserv.linguistlist.org><mailto:Lingtyp at listserv.linguistlist.org
>>
>>     <mailto:Lingtyp at listserv.linguistlist.org>>
>>     >http://listserv.linguistlist.org/mailman/listinfo/lingtyp
>>     > _______________________________________________
>>     > Lingtyp mailing list
>>     > Lingtyp at listserv.linguistlist.org
>>    
>> <mailto:Lingtyp at listserv.linguistlist.org><mailto:Lingtyp at listserv.linguistlist.org
>>
>>     <mailto:Lingtyp at listserv.linguistlist.org>>
>>     > http://listserv.linguistlist.org/mailman/listinfo/lingtyp
>>     >
>>     >
>>     >
>>     >
>>     > _______________________________________________
>>     > Lingtyp mailing list
>>     > Lingtyp at listserv.linguistlist.org
>>     <mailto:Lingtyp at listserv.linguistlist.org>
>>     > http://listserv.linguistlist.org/mailman/listinfo/lingtyp
>>     >
>>     _______________________________________________
>>     Lingtyp mailing list
>>     Lingtyp at listserv.linguistlist.org
>>     <mailto:Lingtyp at listserv.linguistlist.org>
>>     http://listserv.linguistlist.org/mailman/listinfo/lingtyp
>>
>>
>>
>>
>> _______________________________________________
>> Lingtyp mailing list
>> Lingtyp at listserv.linguistlist.org
>> http://listserv.linguistlist.org/mailman/listinfo/lingtyp
> 
> 
> 
> _______________________________________________
> Lingtyp mailing list
> Lingtyp at listserv.linguistlist.org
> http://listserv.linguistlist.org/mailman/listinfo/lingtyp
> 




More information about the Lingtyp mailing list