[Lingtyp] Resources on glossing choices

Mattis List mattis.list at lingpy.org
Tue Dec 31 17:15:27 UTC 2019

Dear all,

as part of the CLDF initiative for cross-linguistic data standards 
(https://cldf.clld.org), we have recently tried to elaborate to which 
degree we could apply the tools we already have (Concepticon, the idea 
of a Grammaticon, CLTS, Glottolog) to try and make inter-linear-glossed 
text more comparable BEYOND one grammar for one particular language.

We have a draft that is almost accepted for publication devoted to this 
topics: https://doi.org/10.17613/gscz-mb13

Comments are welcome, as we're currently revising it to meet final 
requirements by reviewers.

The basic idea is: We all should work on making our work more comparable 
on the g(eneral)-linguistics level (to borrow a term from Martin 
Haspelmath), while nothing prevents us from being consistent on the 
p(articular)-linguistics level.

All the best,


On 31.12.19 17:26, Daniel Ross wrote:
> A quick addition: please do include clear examples of the range of forms 
> when discussing labels/function. This is less often a problem with 
> current grammars, but sometimes it can still be hard to be certain 
> exactly which morpheme(s) correspond to the explanation, and especially 
> complicated in cases of allomorphy. Sometimes this seems to be spread 
> over three chapters: a discussion of allomorphy under Phonology, a 
> discussion of function under Morphology, and examples in usage under 
> Syntax. The result is that a careful reader must cross-reference the 
> forms found in examples in Syntax with the allomorphy in Phonology to 
> link up functions in Morphology, which is hard to do with confidence!
> On Tue, Dec 31, 2019 at 8:20 AM Daniel Ross <djross3 at gmail.com 
> <mailto:djross3 at gmail.com>> wrote:
>     Dear Kate,
>     I'm glad you're asking about this, because this is an important
>     factor in the readability of grammars. As a frequent reader of
>     grammars, I often find myself needing to "decode" glosses (or worse,
>     of course, unglossed forms). From my perspective, the most important
>     consideration is identifiability: if I am quoting an example from a
>     grammar, I want to be able to identify each part, including
>     multi-functionality of morphemes. That allows me to compare them to
>     similar forms in other languages (including comparing
>     multi-functionality of similar morphemes) and also, if needed, adapt
>     the glosses.
>     There are two obstacles that I often encounter in that task:
>     1. Obscure or idiosyncratic morpheme/category labels that make me
>     uncertain about identifying a form.
>     2. Inconsistent glossing such that the (potentially) same morpheme
>     is glossed differently and I'm unsure about drawing a connection.
>     I would think that problem #2 is more significant, although both of
>     these can be solved by having a clear explanation somewhere in the
>     grammar: for example, use a single abbreviation for the form (that's
>     somewhat arbitrary), then have a list of abbreviations at the
>     beginning, with a label clearly corresponding to some section in the
>     table of contents. From my perspective, if it's hard to understand a
>     gloss to begin with, it's probably then necessary to read some
>     background (e.g., how tense marking works in the language) to be
>     confident in interpreting and quoting the example.
>     Another consideration is that I often skim grammars, if they are
>     available as PDFs, by searching for the morpheme form and/or the
>     gloss. If the gloss changes in context that would be confusing, but
>     similarly if there is a lot of allomorphy then searching by form
>     might not be possible either.
>     Regardless, the most important point is as you stated to be
>     consistent*, and if anything is not transparent to explain it
>     clearly. That should be sufficient. Explicit comments on uncertain
>     cases of multi-functionality vs. homophony are also useful.
>     If the glossing is complex enough (and for some languages it must
>     be), then you could include an explicit section in your
>     introduction: I rarely see this, but would appreciate it. Sometimes
>     I do see footnotes when potentially confusing glosses are presented,
>     but those can be hard to locate (or even know to look for) when
>     reading elsewhere in the grammar.
>     [*By "consistent" here I mean using a principled methodology, not
>     necessarily having one-to-one pairings of morphemes to labels,
>     although if possible that seems helpful too.]
>     Daniel
>     On Tue, Dec 31, 2019 at 7:35 AM Christian Lehmann
>     <christian.lehmann at uni-erfurt.de
>     <mailto:christian.lehmann at uni-erfurt.de>> wrote:
>         Dear Kate,
>         the best system of glossing that I know of is on my website (:-)):
>         https://www.christianlehmann.eu/ling/ling_meth/ling_description/grammaticography/gloss/index.php
>         which expands on an earlier publication referenced there.
>         As for your particular problem, I agree with you that the main
>         concern
>         is consistency. Assume you can distinguish between homonymy and
>         polysemy, then even enumerating all the polysemous senses of a
>         morph(eme) may be much to cumbersome for the gloss to be
>         helpful. In
>         such a case, my recommendation would be to treat the set of
>         functions of
>         each morpheme in the respective chapter of your grammar,
>         relieving thus
>         the interlinear gloss of this task, and adding a note that your
>         glosses
>         are, in this respect, context-dependent. To be sure, this is
>         against a
>         more general principle of consistent glossing. However, interlinear
>         glosses serve a practical, not a theoretical function.
>         Best,
>         Christian
>         -- 
>         Prof. em. Dr. Christian Lehmann
>         Rudolfstr. 4
>         D - 99092 Erfurt
>         Tel.:   +49/361/2113417
>         Fax:    +49/361/2113418
>         E-Post: christianw_lehmann at arcor.de
>         <mailto:christianw_lehmann at arcor.de>
>         https://www.christianlehmann.eu
>         _______________________________________________
>         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

More information about the Lingtyp mailing list