On Tue, Feb 7, 2017 at 2:20 PM, Amir E. Aharoni
<amir.aharoni(a)mail.huji.ac.il> wrote:
2017-02-07 15:08 GMT+02:00 Michael Everson
<everson(a)evertype.com>om>:
On 6 Feb 2017, at 21:45, Milos Rancic
<millosh(a)gmail.com> wrote:
I am in favor of making LangCom a normal
democratic body: 50%+1 (of
those who voted) for regular decisions, 2/3 majority (of those who
voted) for changing the rules. ("Of those who voted" because we have u
number of inactive members.)
I would support this.
In general, I'm not a fan of voting about issues in Wikipedia. Democracy is
good for countries, but not necessarily for an encyclopedia. The English
Wikipedia has a pretty clear practice of not deciding about pretty much
anything by vote count. My home wiki the Hebrew Wikipedia is quite
different, and a lot of things are decided by a vote there; I consider it
wrong, and never participate in such votes.
For Langcom, a voting policy will possibly make sense for areas where there
is space for opinion, such as a prediction of a project's viability,
perceived incubator activity, new members, or whether to approve a
constructed language. These things are hard to measure precisely. Also, for
questions such as whether to approve a project with a macro-language code or
not. At the moment it's possible that one opposing committee member will
block progress without even having to explain their opinion, and this is not
great.
Amir, Language committee is not a Wikipedia :P We could just fool
ourselves by saying that we are not making political decisions [based
on linguistic facts].
There are a couple of paradigms because we are making political decisions.
1. The first is related to the structural political influence over the
standardization bodies. Language is an important part of national
identity => national institutions won't recognize some obvious facts
=> international standardization institutions would either
automatically listen to the national institutions or they will be
pressured to make a decision which particular national institution
would prefer.
We don't need to move into the hotter areas: just the recent case of
Elfdalian in liberal Sweden could be paradigmatic for that type of
decisions.
By making a decision contrary or harmonized with the position of a
national (or international) institution, we are making political
decision.
But this type of issues is the simplest one. In this case we are
making the best expert decision, based on linguistic facts. However,
no matter if our decision has good basis in scientific consensus, we
are making a political decision anyway.
2. The second one is related to the scientific consensus itself.
Unlike in, let's say, thermodynamics scientific consensus in marking
something a language or not is significantly tinier even without
structural political influence over the international bodies.
Let's take as an example Romany languages. We have two serious issues
to solve in this case.
2.1. The first one is related to the actual knowledge we have about
Romany languages. SIL/Ethnologue have totally different categorization
of Romany languages from at least the point of former Yugoslavian and
Bulgarian linguistics. (I felt like I prepared wrong lectures for exam
when I went to the main Romany linguist in Serbia with
Ethnologue-based knowledge.)
Note that we are talking in this case about the following languages,
according to SIL/Ethnologue:
1) Romany "macrolanguage" [1]
2) Balkan Romany [2]
3) Baltic Romany [3]
4) Carpathian Romany [4]
5) Kalo Finnish Romany [5]
6) Sinte Romany [6]
7) Vlax Romany [7]
8) Welsh Romany [8]
9) Lomavren [9]
10) Erromintxela [10]
11) Traveller Dannish [11]
12) Angloromani [12]
13) Romano-Greek [13]
14) Calo [14]
15) Norwegian Traveller [15]
16) Romano-Serbian [16]
17) Tavringer Romani [17]
I know that common Roma population of Belgrade (speakers of either
Balkan Romany or "Romano-Serbian", according to Ethnologue) treat as
"close others" Vlax Romani (although not in linguistic sense) and "not
us" likely Sinte Romany population.
However, according to Serbian (and former Yugoslavian) linguistics,
there is *one* Romany language and there is definitely no mixed
language called "Romano-Serbian".
We have Vlax Romany Wikipedia [18] (with 582 articles; created before
Language committee) and request [19] / Incubator [20] for Wikipedia in
Balkan Romany, which has been made eligible in 2007.
*Any* decision that we make here is or would be political. Besides the
fact that we have no clue what should be actually done. For example,
is it reasonable to treat them as one language and insist on one
Wikipedia? The main split from Balkans happened in 15th century,
meaning that the most of the varieties should be closer than English
and Scots are (12th century).
It is normal that we have elaborated and educated different opinions.
It is normal that any of us has a strong opinion in one way: For
example, Oliver could strongly support SIL/Ethnologue, I could
strongly support local linguistics. It's simply irrational to wait
either Oliver or me to change the opinion and unblock particular set
of possible Wikipedias.
2.2. Standardization
Roma people haven't passed through the process of creation of the
national identity. They are doing that now. And, as we know, language
is important part of the national identity. And not any language, but
standardized language, taught in schools all over the areas where the
population is present.
However, postmodern approach is that the children should learn in
their native language [variety], which could be different enough from
the standard and could be treated as a foreign language (obviously,
not as distant as Swiss German vs. Standard German, but distant enough
to be treated as a non-native variety).
Do Roma people need 17 Wikipedias or just one? OK, not 17 -- some of
the language varieties are spoken by few dozens or few hundreds of
people --, but 2, 5, 10?
Again, any decision that we make is a political one. All reasonable
approaches have advantages and disadvantages. Would you support
children to learn in their native language and immediately enjoy
positive consequences of having basic education in native language or
standardization process, which would give to Roma in the future more
political power?
That's again political decision. I could have a strong position
towards one approach, you could have a strong position towards another
approach; and both approaches are quite valid ones. Just because we
have consensus-based decision-making, we could block all the
approaches in Wikimedia environment and that's the worst thing we
could do.
3. The third major reason is related to the significance of Wikipedia
in contemporary civilization. By making decision to do one thing or
another, we are quite likely giving significant advantage to our
preferred option.
Here is one hot topic to illustrate real political, even
life-and-death consequences of our decision.
Zaza is Kurdish subethnicity, which is in the process of separating
themselves from Kurds. The process is at the beginning. It could
become more influential or it could vanish in relatively near future.
Zaza is a language of different branch of Northwestern I languages
from Kurdish (and possibly related to the Caspian languages).
But they were not at all linguistically oppressed inside of the
Kurdish population. Kurdish newspapers are bilingual: in Kurdish and
Zaza. Kurds even standardized Zaza language. (This article is good to
read for the context [21].)
However, during the last 10-20 years, there is a concurrent
institution called "Zazaki Institute" [22], which built a separate
standard. We have a low level conflict on Meta and Translatewiki which
lasts for a decade (start here [23][24]).
Sooner or later we will have to make our decision, without
bureaucratic excuses. And that decision is going to be very political.
The options are: (1) De facto support Kurdish national unity and
significantly influence likely valid right to self-determination of
one ethnicity. (2) De facto support self-determination right and
likely position ourselves on the line with Turkish intelligence
agencies.
No matter what we decide, it will be political. We could have
different opinions in relation to the Kurdish right to
self-determination (in relation to Turkey, Syria, Iraq, Iran) and we
could have different opinions in relation to the Zaza right to
self-determination (in relation to Kurds).
It is obvious that we would have to think much more deeply about this
issue than about the standardization. However, political decisions
should be made politically, not bureaucratically. There is the real
world outside, our decisions influence it and we should stop avoiding
that responsibility.
For more clear-cut questions such as whether to
approve a project in a
language without an ISO code, there shouldn't be a vote—it should be an
immediate rejection.
BCP 47 is often the only reasonable way to make something working. For
example, although we didn't make decision in relation to that, but
Belarusian Tarashkevitsa is a legitimate written variety and there is
no easy translation engine between the official standard and
Tarashkevitsa. (Logic goes: Standard Belarusian is official in
Belarus; however, Russian is the native language for the most of those
who use standard; majority of those who actually use Belarusian use
Tarashkevitsa.)
Take a look into the list of Quechuan languages [25]. It's a primary
group (according to the traditional, non-Greenberg categorization) of
44 languages (according to SIL/Ethnologue). Those 44 languages are
spoken by ~8 million people.
But not just that. At least Ecuador is working on one, standardized
language to be used by Quechua population in Ecuador. And unlike the
relation between Chile and Mapuche, it seems that the indigenous
population has positive attitude to the standardization.
What would be the code if they come and ask us for Wikipedia? Should
we choose a code for a random Quechua language spoken in Ecuador?
Should we wait for JAC, which could take years? Isn't it much more
reasonable to give them BCP 47 code "qu-ec" and change it when JAC
standardize it?
Any language which have two scripts and population. North and South
Azerbaijani division is a joke. If blocked Turkish Wikipedians could
go to Azerbaijani Wikipedia and edit there, then the differences
between Azerbaijani spoken varieties are quite small (an average
Serbian Wikipedian wouldn't be able to go to Macedonian, Bulgarian or
Slovenian Wikipedia). We were just lucky that there is a separate
code, which could be used for the other written variety (although I am
still of opinion that one Azerbaijani Wikipedia is enough). The right
code for Azerbaijani variety written in Arabic script is "az-arab",
not "azb".
I am sure it's similar with the difference between Central and
Peripheral Mongolian. (But there are significantly less troubles in
switching from left-to-right to right-to-left in comparison to
switching from left-to-right to top-to-down.) We could use wrong code
for Mongolian written in Mongolian script, if it's a matter of would
we have that project or not, but the right code for that project is
"mn-mong".
4. We have the rules, which could and should be changed from time to
time. What we want is also political and we should have the process
which would include everybody's positions. But, eventually, we
shouldn't depend on ability of one person -- even it could be myself
-- to block the changes about everybody else agree.
5. Fortunately for Affiliations committee, it is not the most
dysfunctional Wikimedia committee as long as Language committee exists
under the rule of consensus. There are many things which we missed
because it's either not possible to do or it costs too much efforts.
6. While inactive members do not make any trouble, I don't think that
we should create the House of Lords here. If somebody is not able to
give minimum inputs for approximately the time of LangCom's existence,
something is wrong with our structure. And it's not about deciding who
will be a member of LangCom or not, but about creating rules that
would be applied for everybody.
[1]
https://www.ethnologue.com/language/rom
[2]
https://www.ethnologue.com/language/rmn
[3]
https://www.ethnologue.com/language/rml
[4]
https://www.ethnologue.com/language/rmc
[5]
https://www.ethnologue.com/language/rmf
[6]
https://www.ethnologue.com/language/rmo
[7]
https://www.ethnologue.com/language/rmy
[8]
https://www.ethnologue.com/language/rmw
[9]
https://www.ethnologue.com/language/rmi
[10]
https://www.ethnologue.com/language/emx
[11]
https://www.ethnologue.com/language/rmd
[12]
https://www.ethnologue.com/language/rme
[13]
https://www.ethnologue.com/language/rge
[14]
https://www.ethnologue.com/language/rmq
[15]
https://www.ethnologue.com/language/rmg
[16]
https://www.ethnologue.com/language/rsb
[17]
https://www.ethnologue.com/language/rmu
[18]
https://rmy.wikipedia.org/
[19]
https://meta.wikimedia.org/wiki/Requests_for_new_languages/Wikipedia_Balkan…
[20]
https://incubator.wikimedia.org/wiki/Wp/rmn
[21]
http://www.zazaki.net/yazdir/haber/the-zazas-a-kurdish-sub-ethnic-group-or-…
[22]
http://zazaki-institut.de/
[23]
https://meta.wikimedia.org/wiki/Requests_for_comment/Zazaki_wikipedia
[24]
https://meta.wikimedia.org/wiki/Requests_for_new_languages/Wikipedia_Standa…
[25]
http://www-01.sil.org/iso639-3/documentation.asp?id=que