We are planning a new conference on computational modeling in linguistics for Fall 2017 and would love to get input from phonologists and other linguists about how we can best realize our goals for this conference and make it as accessible and useful to linguists as possible.

Please follow the link to our lab blog for full details and discussion, but to summarize, this conference is intended for linguists and cognitive scientists using computational and mathematical approaches to study the human language faculty.  We’re particularly interested in input from linguists on the following key aspects of the current plans:

• The short-term and long-term co-location/venue possibilities
• The name and ideal scope/audience for the conference
• The tentative plan to have short paper (6-8pp) submissions (rather than abstract submissions). Please go to this post for discussion and comments on this aspect of the current plans.

We’ve had quite a bit of discussion on our blog already, primarily from potential participants of this conference who also regularly attend workshops and/or conferences affiliated with the ACL. This discussion has been very useful, but the conference would ideally be accessible and appealing to a broader community of linguists than is currently represented in the discussion. So we are looking for additional input from linguists who are potentially interested in this new conference, but who do not generally attend ACL – we’d be very grateful to hear about the kinds of considerations that could increase or decrease the chances phonologists and other linguists would attend and/or submit their best work to this conference.

Please visit our lab blog to read the details and to leave your comments!

Many thanks!
Gaja Jarosz

We’re working on a statistical meta-analysis of studies investigating phonetic naturalness bias in artificial grammar learning, and we’re looking for unpublished studies (to attempt to correct for publication bias).  If you or your students have such a study, we would be most grateful if you could send us the manuscript or handout (with information about the design, and enough data analysis that we can calculate standard effect size measures).  We’re looking for both adult studies and child studies, as well as studies that do and do not control for structural complexity.  In general, we’d rather receive studies that need to be excluded based on the criteria of our meta-analysis than miss studies that could have been included.

We would greatly appreciate if you could send your studies to wendell.kimper@manchester.ac.uk before 1st October 2016.

Thanks so much!
Wendell Kimper and Anna Greenwood

From Joe Pater

I was reading Syntactic Structures recently for a non-phonological project, and I was surprised to come across in a footnote an analysis of an irregular past tense alternation that looked a lot more like Albright and Hayes (2003) (and maybe Rumelhart and McClelland 1986) than Pinker and Prince (1988) (the footnote is on pp. 58-59 – see also the citation of Hockett in the full footnote):

Has anyone seen this mentioned anywhere else? Albright and Hayes and Pinker and Prince don’t talk about it, and I assume Rumelhart and McClelland don’t either. Other discussion is of course welcome…

From Shigeto Kawahara

Dear colleagues,

I am looking for examples of the following type. In loanword adaptation, sometimes the borrower start using a structure that was not allowed in their native language. For example, Japanese voiced geminates were not allowed in native words, but nevertheless Japanese speakers started using voiced geminates in loanwords. I am looking for examples in which words with these structures are “nativized” and consequently are not allowed to have that structure any longer. For the Japanese case, when a loanword containing a voiced geminate becomes familiar enough (“nativized”), that voiced geminate is eliminated.

If anybody knows other examples of this kind, I’d love to know.

Thanks in advance,

Shigeto

From Joe Pater

I came across an interesting blog post the other day discussing the practice of posting conference papers to arXiv in NLP and machine learning before they have been reviewed. It includes some data from a poll on how people use it in each discipline – machine learning people tend to post earlier in the publication cycle, perhaps due to an influential call for a new publishing model by Yann Le Cun of deep learning fame, and perhaps due to a greater fear of being scooped.

This got me thinking again about archives in our discipline. I came of academic age at the time that ROA was launched, and it was fantastic as a grad student to have access to the latest research in the framework I was using, and to be able to share my own work so easily. As I’ve told Alan Prince already, we’re hugely in his debt for having established that archive, and we also owe a huge thanks to Eric Bakovi? and others for all their work on it, as we do to Michal Starke and others at LingBuzz.

It’s clear, though, that in contrast with the situation in computer science, use of archives is on the decline in phonology. I post to them only sporadically myself, generally only making time to keep my own web page updated. In contrast to when ROA was founded, the preservation function of an archive is less required; most of us have archives serving this purpose at our own institutions (see e.g. John McCarthy’s ScholarWorks archive), and who knows, maybe a document hosted on a google drive will last longer than one on a university site. I find the google drive alternative particularly convenient because it’s so easy to update a paper. And this brings up the main issue in my mind for posting to archives early in the publication cycle: if you have your paper in multiple places, you need to update multiple copies, each with considerably more hassle than a google drive.

Preservation is only one function of these archives, and it’s far less important than another: dissemination. For dissemination, one’s own webpage, or institutional archive, is not a viable alternative. The main impetus for Phonolist was to facilitate dissemination for papers that weren’t being posted to the archives, and it seemed that the added functionality of optional blog discussion of papers would make it attractive for that purpose. I’ve been somewhat surprised to see that people haven’t been using it much for that (most of the papers we advertise are reposts from LingBuzz and ROA).

Phonolist currently lacks any indexing functionality (besides searches), and this is one way that it could be improved to better serve the cause of dissemination. This will likely be an upcoming addition, along with a community .bibtex file.

The question I’d like to bring up for discussion is whether people perceive the need for a general phonology archive, and if so, what it should look like. ROA is limited to OT and its affiliates, and LingBuzz has technical issues that have made it frustrating to use, and I’ve heard that it’s unlikely to be improved. My limited experience with academia.edu and researchgate has been negative. I thought an easy fix might be to start using http://cogprints.org, but in response to my inquiry about it, Stevan Harnad said “CogPrints has no long-term support and I would say it’s obsolete (though I’m still keeping it up).” More generally, I’d be interested to hear people’s thoughts about how they use the existing archives, and why they don’t use them.

From Stephanie Shih (email: shih@ucmerced.edu)

Spurred on by the interesting discussion of phonology at general linguistics conferences (initiated by Abby Kaplan, https://websites.umass.edu/phonolist/2016/05/31/discussion-phonology-at-general-linguistics-conferences/), I was curious what the actual numbers are.

Is it the case that the data reflect our intuitions about the footprint of phonology? Here is the data, between 1978 and the present for five general conferences: GLOW, BLS, CLS, NELS, and WCCFL. Details of data collection are included at the end. Raw data is available here: http://stephsus.github.io/linguometrics

Overall, the average representation of “p” papers at general linguistics conferences sits at between 24–27% for the time period, 1980–2016. GLOW is a major outlier, at a historical 13%. WCCFL has the highest historical average, at 27.24%, though CLS matches this in the recent data that is available (average 27.5%, for years 2009–2016). So in terms of getting the most “bang for your buck” if one wants to hobnob with the most phonologists, CLS looks like the best recent bet. In comparison to these averages, the % dissertations on the “p” side from 2009–2012 is ~19% (from this Proquest count: http://languagelog.ldc.upenn.edu/nll/?p=4349). This 19% includes both phonetics and phonology dissertations. Counting just phonology: 13.59%.

I think there’s been a general intuition that I’ve heard echoed both in the Phonolist discussion and in private conversations that phonology is increasingly less well-represented at general linguistics conferences overall. Looking at the data, there certainly is a marked drop between 2009–2013, for NELS, WCCFL, and even BLS. I’m quite curious what caused this drop. 2013 was the first year of AMP, but that is probably more of a reaction to the perceived drop in phonology representation than the cause of it. In 2014–2015, there looks to have been a recovery of “p” papers at the various conferences, even though there are more “p”-specialty conferences available now (e.g., AMP, LabPhon, etc.).

Also from eyeballing the data, it doesn’t look like the introduction of AMP has lessened the number of “p” papers at other general conferences, at least in the short time period that AMP has been around so far. In both 2014–2015, levels for most of the other conferences rise. There’s been a bit of a drop in 2016 so far, and while this might be due to the increasing popularity of AMP, it’s probably more reasonable to suspect that it’s just the vicissitudes of a natural research cycle, as we can see from the previous 30 years.

Ultimately, the answer of what causes fluctuations in the “p” participation in conferences is most likely a multi-factor one. The temporal closeness of AMP and NELS, for instance, might in the end bleed off some NELS participation (but affect, say, WCCFL less). Our perception of trends probably also affects submissions (e.g., “oh, Conference X didn’t have a lot of p-papers last year, so we won’t submit there this year,” or “oh crap, we should submit to Conference X now to keep up p-representation”). Additionally, who the hosts are, who the reviewers are, and who the invited speakers are might matter. Juliet Stanton suggests also that if there is a specific phonology session at the conference might encourage “p” participation, so perhaps in addition to an effort to submit individual papers to general conferences, making an effort to organize special phonology sessions at conferences will increase “p” with greater impact.

Data

Data was collected based what was easily available on the web. Most NELS, WCCFL, and BLS numbers are based on the table of contents of the published volumes, and thus may not exactly reflect the actual programs of the conferences. When no published volumes were available, data from web-available conference timetables were used (either from the existing conference website or Linguist List posting).

Years reflect the year in which the conference was held, not the year in which the proceedings were published. No distinction was made between invited talks, regular talks, special session talks, and posters. If anyone has conference programs from the missing years available, I would happily welcome the contribution to fill in the missing data.

Coding

Papers that fall into the “phonetics & phonology” (i.e., “p-oriented”) range were coded based on best educated guess given the title by me (and for the GLOW data, by Juliet Stanton). For papers that were about morphology or the phonology-syntax interface, I erred on the side of inclusive, so many of these are counted for “p”. In particular, morphology and phonology-syntax interface papers whose primary authors identify as phonologists were coded as “p”, under the assumption that  part of having a phonological representation at a conference is about the researchers themselves in addition to actual papers about phonology. If anything, these numbers for “p” are more inclusive than exclusive.

(Acknowledgements to Juliet Stanton for the GLOW data, and to Adam Chong for discussion about this data.)

From Abby Kaplan:

I’d like to solicit thoughts about the representation of phonology at general linguistics conferences.  (I’m thinking particularly of major regional conferences such as WCCFL and NELS.)  As I’m sure we’re all aware, the phonology talks are often substantially outnumbered by syntax talks – sometimes so much so that a phonologist may have only a couple of sessions to attend.

This topic has been on my mind lately because my institution hosted WCCFL 34, so I’ve had a front-row seat to some of the raw numbers.  This year, there were 157 submissions in syntax and 23 in phonology (plus 2 in both); the composition of the final program reflected this trend.  [NB: this is not an official post on behalf of the WCCFL organizing committee, although all the committee members have seen it.]

The situation isn’t new, of course, and I can imagine a few reasons for it.  One is simple demographics: there seem to be more syntacticians than phonologists (as suggested by the data here and in related posts).  Another is the existence of excellent phonology-specific conferences (mfm, LabPhon, AMP, NAPhC, etc.), which may be siphoning off work that otherwise would have gone to a general conference.  I also don’t mean to suggest that the situation is dire.  The difference may be more pronounced at some conferences than others; for example, my impression from the last few years’ programs is that there’s proportionally more phonology at CLS and BLS than at WCCFL and NELS.  And I certainly saw excellent phonology research at this year’s WCCFL.

I have no problem with being in a smaller subfield; I’m not arguing that we should try to catch up with the syntacticians.  And I know that the situation isn’t unique to phonology.  But I’d like to hear other people’s thoughts about what this means for general conferences.  What happens when the numbers get so small that serious questions arise about the value of a particular conference for phonologists?  Is it worth making a special effort to submit to conferences like WCCFL in order to sustain critical mass there?  On the one hand, I can see the value in working to maintain a phonology presence at conferences that are supposed to represent the field pretty broadly.  But on the other hand, if you want to get the best possible feedback on your work and see what other phonologists are doing, then submitting to a conference that’s light on phonology, just for symbolic reasons, may not do you the most good.

Thoughts?