Sender: lojban@googlegroups.com
Received-SPF: neutral (google.com: 35.164.127.236 is neither permitted nor denied by best guess record for domain of lojbab@lojban.org) client-ip=35.164.127.236;
Subject: Re: [lojban] Glossing
To: lojban@googlegroups.com
References: <3439459.Bz73Tnfg3h@mooncat>
 <CAOYwWb2turc=Q1Zi2NzwZATC=HT7u=NCNR4mWzsWyxo_W40ViQ@mail.gmail.com>
 <7a6d7861-7271-8902-bf73-1d134f5b9f07@lojban.org>
 <CAOYwWb2jT49EO87iNeAbQsmdi4mXWRf4eGbU8Csshimy6mB7gg@mail.gmail.com>
From: Bob LeChevalier <lojbab@lojban.org>
Message-ID: <b039b4a2-597e-3978-0b0a-d4e90fc5d5fe@lojban.org>
Date: Fri, 26 Apr 2019 08:45:52 -0400
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:60.0) Gecko/20100101
 Thunderbird/60.6.1
MIME-Version: 1.0
In-Reply-To: <CAOYwWb2jT49EO87iNeAbQsmdi4mXWRf4eGbU8Csshimy6mB7gg@mail.gmail.com>
Content-Type: text/plain; charset="UTF-8"; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: quoted-printable
Reply-To: lojban@googlegroups.com
Precedence: list
Mailing-list: list lojban@googlegroups.com; contact lojban+owners@googlegroups.com
X-Spam_score: -3.0
X-Spam_score_int: -29
X-Spam_bar: ---

I hope my URLS appear correctly here.

On 4/25/2019 9:55 PM, Mike S. wrote:
> On Thu, Apr 25, 2019 at 5:20 PM Bob LeChevalier <lojbab@lojban.org=20
> <mailto:lojbab@lojban.org>> wrote:
>      > i=C2=A0 =C2=A0lo=C2=A0 bicrbombu=C2=A0 cu=C2=A0 se=C2=A0 =C2=A0 =
=C2=A0 =C2=A0bevri=C2=A0 =C2=A0 =C2=A0 mi o'adai            =20
>      =C2=A0lo=C2=A0 mamta=C2=A0 be=C2=A0 mi vau
>      > SEP ART bumblebee? SEP PASS.INV-carried-by-me (pride-EMPATH)
>     [-to] ART mother-OBJ-me TERMIN
>=20
>=20
> I am no expert at glossing (far from it), but I think there are some=20
> possible points of confusion here where things are supposed to line up. =
=20
> I think hyphens should generally indicate morpheme breaks within words. =
=20
> Dots seem to be used when several gloss-components hold together in one=
=20
> object-language morpheme.=C2=A0 So your gloss ought to be something like:
>=20
> SEP ART bumblebee SEP PASS.INV carried.by <http://carried.by> me=20
> (pride-EMPATH) [to] ART mother OBJ me TERMIN

I agree with the dot instead of the hyphen after "carried" (but not the=20
URL insert, which I am pretty sure you did not intend).  And I should=20
have used a dot instead of a hyphen before the "to" for the same reason=20
(or maybe a triple dot omitted.ellipsis since this is continuing a=20
previous paradigm).

>     I think only SEP (separator), TERMIN (terminator) and EMPATH (empathi=
c
>     attitudinal) are not in the standard abbreviations list, and the rest
>     should be pretty obvious.=C2=A0 I also used the bracketed [-to] which=
 is
>     actually part of the gloss of bevri, as symbolized by the hyphen,
>     put in
>     a more readable position (5+ place brivla would be even harder to
>     express without such readability aids).
>=20
>=20
> I like my idea of e.g. VPZ .... VPT because abbreviations like VP are=20
> already widely understood among English-speaking linguists and the=20
> -Z...-T convention (or something like it) can be used to convey the=20
> not-widely-understood functions of these Lojban particles, which IMHO=20
> ought to be one of the points of writing the gloss in the first place.

The latter would seem to be the key issue.  Why are we writing the=20
gloss?  The most usual reason for writing a gloss is to clarify what is=20
going on for the reader.  The reader in this case is almost always a=20
Lojban learner and not an "English-speaking linguist" where the latter=20
word refers to the academic professional field rather than the=20
learner-of-languages.  Most learners don't know the "widely-understood"=20
abbreviations, nor the technical terms they represent.  Looking at the=20
wikipedia list of abbreviations  (=20
<https://en.wikipedia.org/wiki/List_of_glossing_abbreviations> ), I=20
didn't recognize most of the terms they represent, and I did a=20
personal-if-amateur study of Comrie's linguistic universals work early=20
in my Lojban designing effort.  I see no particular advantage in using=20
words that most people don't know, and especially if we are using them=20
in ways that only approximate the technical meaning.

I thus favor for teaching the language, the introduction of standard=20
terminology that is specific to the Lojban design, and is not beholden=20
to academic linguistic norms that learners likely won't know, and=20
academics would be prone to quibble with (and indeed academics DID=20
quibble with a lot of the usages that I made, and even more that JCB=20
made in the original Loglan design).

That is why brivla, bridi, selma'o (and their various capitalized names)=20
are used in Lojban documentation.  We don't need to argue whether a BAI=20
is a modal, or an adverb, or a preposition.  It is serving the Lojban=20
grammatical function of a BAI, which possibly might be any of those=20
terms in some context that an academic linguist might quibble with.  (I=20
am reminded of JCB's example which in Lojban is raumoi "enough.th" as an=20
ordinal number).  Similarly many of the members of UI could be assigned=20
to specific abbreviations on the list of the "-ive" variety, but in=20
studying Lojban probably one should simply view them as UI for=20
grammatical explanation, or possibly one of several categories of UI=20
(which in my baseline cmavo lists are shown as UI1 (attitudinals), UI2=20
(evidentials), UI4 (emotion aspects). UI7 (emotion contours), etc.

Of course when writing for the academic linguistic world, the gloss=20
needs to conform to academic norms such as the "Leipzig rules" for=20
morphemic glossing (see the google search I referenced to find these).=20
But in such cases, I think the person writing the paper should be=20
clarifying the abbreviations being used and what they are actually=20
representing in Lojban that might not apply to other languages.  Makes a=20
lot more sense than a huge chunk of the abbreviations in the standard=20
list that are defined solely by how they are used in a couple of=20
specific languages e,g, "adessive", "antessive".  And then there are=20
terms like "classifier" which Pierre just quibbled on as applying only=20
to how things are counted.  The wikipedia definition is more generic and=20
it would definite include bic- as a classifier morpheme, which is in=20
fact what it was designed to be, after the fashion of the Dyirbal=20
classes described in=20
<https://en.wikipedia.org>/wiki/Women,_Fire,_and_Dangerous_Things,=20
except that Lojban doesn't have "noun classes" (=20
<https://en.wikipedia.org/wiki/Noun_class> ) since it doesn't have nouns=20
- just brivla which might serve as nouns, verbs, adjectives, and in some=20
forms adverbs, and a bunch of other things.  I would argue that=20
"classifier" is the right linguistics term, based on=20
<https://en.wikipedia.org/wiki/Classifier_(linguistics)> which states

"Languages with classifiers may have up to several hundred different=20
classifiers, whereas those with noun classes (or in particular, genders)=20
tend to have a smaller number of classes, not always much dependent on=20
the nouns' meaning, and with a variety of grammatical consequences."

Lojban Type.III.fu'ivla prefixes can be any of the rafsi (affixes) and=20
are highly dependent on the brivla's "meaning" and with no grammatical=20
consequences, and thus "noun class" would be the wrong term even if that=20
was what actually inspired the design element (there were other=20
considerations besides that book that led to this sort of classifier in=20
Type.III.fu'ivla, most especially the efforts in creating words for=20
Linnean terms and chemical element names, both of which JCB had tackled=20
in Loglan with poor solutions, but Lakoff's book was much talked about=20
at the time).

>     I don't sense that morphemic glossing necessarily is as
>     comprehensive in
>     the structure words as Mike S's attempt provides.=C2=A0 After all, En=
glish
>     has several kinds of articles and they all probably gloss as ART.=20
>     There
>     is an abbreviation for "definite" (DEF), but none for "indefinite", t=
wo
>     of the kinds of articles in English.=C2=A0 But probably if we wanted =
to
>     systematize glossing in Lojban we might invent abbreviations to
>     distinguish lo, la, lei, and loi (there is no abbreviation for
>     mass-nouns either in the standard list)
>=20
> You can use dots to add features to a gloss.=C2=A0 French "la" might be=
=20
> glossed as ART.DEF.FEM and Lojban "loi" might be ART.MASS.

It might, *but* I think that would be an incorrect choice.  The use of=20
capitalized terms are supposed to be about referencing grammatical=20
concepts and effects, and not content concepts.  Arguably, all of the=20
members of selma'o LE are simply articles, and which flavor of article=20
has utterly no grammatical effect, UI and BAI and PA would also be=20
grammatical categories that in theory have no grammatical effect, and I=20
only favor sub-categorizing them (with numbers after the selma'o)=20
because in actual usage, attitudinals, evidentials, contours, etc have=20
distinctly different effects on meaning even while having the same=20
nominal grammar, and those sub-categories are useful in explaining=20
actual usage in Lojban to a learner (or perhaps a linguist), whereas the=20
various "-ive" linguistic terms in the standard list that might be=20
applied only to individual cmavo would imply grammatical functionality=20
that isn't part of the language (e.g. "fa'a" which I think corresponds=20
to "adessive or venitive case" but in fact has nothing to do with=20
linguistic "case", and is two linguistic terms covering something that=20
is only one cmavo in a set having real grammatical value (selma'o FAhA),=20
because they combine in a grammatically predictable way with MOhI, where=20
I suspect there are no standard terms for most of the other members of=20
FAhA that have identical grammar but are not adessive, and nothing so=20
far as I know that can describe MOhI).

Loglan/Lojban was designed by intent to be extreme linguistically with=20
respect to many language norms that are mandatory in some languages=20
while being ignored in other languages.  People know about number and=20
tense as being important parts of Lojban's elimination of mandatory=20
features, but equally important are these other lesser known features=20
that are also non-mandatory, but are permitted (even though they may=20
have no clear correspondence in the speakers native language, hence=20
things like our aorist-like tenses (ZAhO), evidentials, etc.)

How much a language speaker comes to use these categories in Lojban that=20
are not used in their native languages, and to what effect, is clearly=20
part of the realm implied by the Sapir-Whorf testing aspect of=20
Loglan/Lojban.  I have frequently said that Lojban's enormous and highly=20
flexible attitudinal system might show more in the way of Sapir-Whorf=20
effects than the features of the original Loglan design (formal logical=20
connectives, logical predicate brivla, uniquely parsible "metaphoric"=20
modification).  TLI Loglan lacks that enormity and flexibility in its=20
corresponding set.

And the features that might make a difference in forms of usage are the=20
ones linguists SHOULD be interested in when trying to understand how=20
Lojban grammar works.  For that reason, the use of selma'o names with=20
subcategory numbers where applicable is far better than the use of=20
standard terminology, and also has the benefit of being more=20
understandable to the layperson who doesn't know the linguistic jargon.

On the other hand, I once proposed to some academic linguists the use of=20
Lojban as itself a form of linguistic jargon interlanguage for conveying=20
glosses of other languages and respecting their features.  Lojban in=20
effect has a superset of most of the features of pretty much all other=20
languages, even if they aren't conveyed via mandatory grammatical=20
features in Lojban.  Thus Lojban can convey the complexity of a Nootka=20
sentence/word as a Lojban tanru or even a many-part lujvo, as well as=20
the variety of cases in the Finno-Ugric languages.  Using the terms=20
invented to describe these features in academic descriptions of those=20
languages limits one to an audience of those who know the language=20
enough to understand the terms.  Lojban as an interlanguage loses=20
information about what is conveyed via grammar words vs via content=20
words in the source language, but might be more effective in conveying=20
the semantics that result.

Enough for now.  I've essentially denigrated the purpose of Pierre's=20
original question, which was not what I started trying to do in=20
answering him.  But it has been a long while since a question came up=20
here that strongly ties back to the original design concepts of the=20
language, as well as Loglan/Lojban's ties, both positive and negative,=20
to the academic linguistic community (which for the most part haven't=20
had much respect for artificial languages, though I'm not in tune enough=20
to know whether the attitudes of academic linguistics might have changed=20
since the 80s when Lojban was designed).

lojbab

--=20
You received this message because you are subscribed to the Google Groups "=
lojban" group.
To unsubscribe from this group and stop receiving emails from it, send an e=
mail to lojban+unsubscribe@googlegroups.com.
To post to this group, send email to lojban@googlegroups.com.
Visit this group at https://groups.google.com/group/lojban.
For more options, visit https://groups.google.com/d/optout.