« first day (502 days earlier)      last day (4311 days later) » 

02:38
0
Q: Is there an agreed upon feature set that represents vowels?

dmonopolyI have a feeling the answer is no, and that there are complications involved, but I was considering this: [-consonantal, +syllabic] This would first remove all consonants, leaving e, u, i, and others, and then ensure that the remaining are syllabic, which would rule out glides (like j, w). All ...

 
3 hours later…
05:16
anyone up for a chat?
 
6 hours later…
10:52
0
Q: Are English homonyms distinguishable by pitch profile?

timquinnI was told years ago by a teacher at a Carden School that they teach their students that English homonyms, especially those with diphthongs, can be told apart by the pitch profile of the vowel sound. Now I see in wiki that all english diphthongs have a falling tone. This does not necessarily mean...

 
2 hours later…
12:33
@hippietrail Hi, I haven't seen you in a while!
@OtavioMacedo: hi!
In which part of the world are you now? :-)
i've been back in australia for almost three months now
i'm mostly active on travel.SE which just graduated a few days ago
12:50
Oh, that's good news!
I knew it would
It was doing very good.
13:03
yeah we're pretty happy, and a bit of an influx has begun
i was surprised how slow linguistics turned out to be
Yeah, we're considering "raising the bar"
You know, only accept high-level research questions.
the problem is to attract them
i think we lowered the bar months ago because there were not enough questions at all
Exactly. There is a fine balance to be struck.
when we first started the first thing i noticed was so many what i considered crap questions
and a few that were way out of my league even to comprehend (-:
Yes, that's another problem.
We need at least one professional linguist as moderator.
But they don't seem to be much interested.
13:11
perhaps we could try leveraging the NLP angle from stackoverflow
since we're not attracting linguists from out of the blue
Hmm tell me more about that
well it's traditionally proven to be difficult to attract contributors from nowhere - mostly the wander across from stackoverflow
and of course the two areas where linguistics intersects with programming are natural language processing and computational linguistics
2
so if there were a way to encourage stackoverflow people to ask good practical questions in those fields it would bump our good question ratio
then when we have a higher good question ratio in at least some fields it could make us appear more credible to linguists generically, at first to ones that work in NLP/CL as well as other fields
Interesting! I'll start monitoring stackoverflow for NLP questions, then.
13:45
i'm interested in both fields myself, though i never actually do any nlp programming other than my ancient open source machine translator project i haven't touched in years
14:12
You might be interested in this: coursera.org/course/nlangp

« first day (502 days earlier)      last day (4311 days later) »