Language Log

Below is a guest post by Andrew Caines:

There's been growing interest in recent years in crowdsourcing as a means of data collection: for example, asking workers on Amazon Mechanical Turk to rate sentences for grammaticality, implicatures, sentiment, etc. As part of a special session for this year's INTERSPEECH Conference on innovative uses of crowdsourcing, we're building a crowdsourced spoken corpus of English and German.

James Madison, "Notes on the Debates in the Federal Convention", Tuesday September 17, 1787:

Whilst the last members [of the Constitutional Convention] were signing [the final document], Doctr. FRANKLIN looking towards the Presidents Chair, at the back of which a rising sun happened to be painted, observed to a few members near him, that Painters had found it difficult to distinguish in their art a rising from a setting sun.

I have, said he, often and often in the course of the Session, and the vicissitudes of my hopes and fears as to its issue, looked at that behind the President without being able to tell whether it was rising or setting: But now at length I have the happiness to know that it is a rising and not a setting Sun.

I'll be spending next Friday and Saturday in San Francisco at Text By The Bay, billed as "A new NLP conference bringing together researchers and practitioners, using computational linguistics and text mining to build new companies through understanding and meaning."

With 46 interesting-looking talks and a couple of panels, this seems like an excellent way to get a sense of the opportunities and activities in this area. There are talks from people at Microsoft, Wikimedia, AirBnB, Trulia, Ancestry.com, Bloomberg, OpenTable, Twitter, LinkedIn, Verizon, etc., and from people at Berkeley, Stanford, Penn, and Purdue. And some of the presentations by people from smaller, newer, less-familiar outfits may be the most interesting of all.

Registration is not cheap — "new companies", an expensive venue, and all — but the organizer, Alexy Khrabrov, tells me that the discount code TEXTMARK will get you 50% off, and students who email to text@bythebay.io from their university account may be able to negotiate further reductions.

Two years ago I sent out an "SOS for DARE," that is, a plea for the indispensable Dictionary of American Regional English, which had run into funding troubles. Though DARE was granted a temporary reprieve, the latest news is more dire than ever.

Marc Johnson laid out the situation in an article for the Milwaukee Journal-Sentinel:

The end may be near for one of the University of Wisconsin-Madison's most celebrated humanities projects, the half-century-old Dictionary of American Regional English. In a few months, the budget pool will drain to a puddle. Layoff notices have been sent, eulogies composed…

Tomorrow afternoon, I'll be giving a talk at the Simons Foundation (160 5th Avenue, New York NY) with the title "Reproducible Research and the Common Task Method".

Despite the April 1 date, the topic is a serious one. For some background on why the concept of "Reproducible Research" is currently a hot topic, see Paul Voosen, "Amid a Sea of False Findings, the NIH Tries Reform", Chronicle of Higher Education, 3/16/2105:

While the public remains relatively unaware of the problem, it is now a truism in the scientific establishment that many preclinical biomedical studies, when subjected to additional scrutiny, turn out to be false. Many researchers believe that if scientists set out to reproduce preclinical work published over the past decade, a majority would fail. This, in short, is the reproducibility crisis.

Below is a guest post by Marten van der Meulen, who is a teaching assistant for this course.

On March 30th, the Massive Open Online Course (or MOOC) Miracles of Human Language: an Introduction to Linguistics will start on Coursera. The course is facilitated by Leiden University, and is given by Marc van Oostendorp, professor at Leiden University and the Meertens Institute. Subscribing is still possible.

From Christie Versagli:

It's with enthusiasm that we at the World Well-Being Project (University of Pennsylvania) would like to share with you the launch of lexhub.org, a hub for data, tools, papers, and almost any resource in the growing field of language analysis for social science. 

Yesterday afternoon, a popular link from the Washington Post (Ana Swanson, "The equivalent of “It’s all Greek to me” in 30 other languages", Wonkblog 3/25/2015) caused a spike in LLOG page views; this happened to cause a disk drive to fill up, because the back-end database server was keeping binary logs of all transactions; this caused and/or uncovered various other problems; and so LLOG was down for about 24 hours.

More specifically, the site displayed

Error establishing a database connection

in response to nearly all attempts to display WordPress pages.

As a result of several hours of intelligent and heroic labor by Wayne Hill, we're back, with updated and better-configured version of all the underlying software packages. So performance should be better, but in any case, things are working.

Earlier today, someone set up a Facebook account https://www.facebook.com/mark.liberman.121, with a version of my FB profile picture, and began communicating with people as if they were me. My actual FB page is https://www.facebook.com/mark.liberman, which I don't use much except to look at things that people tell me about.

This is apparently a phishing enterprise. The impostor asks people for their phone number and email address and postal address — at least one person gave them this much information — and eventually gets around to money.

A year ago, I wrote "Zhou Youguang, Father of Pinyin" (1/14/14) to celebrate Zhou xiansheng's 108th birthday and his many accomplishments in language reform and applied linguistics.  Included in that post were a portrait of ZYG in his study and numerous links concerning the man and his works.

There's a new linguablog that's definitely worth your time if you're not put off by vulgarities. And if you revel in vulgarities, well, you're in luck. It's called Strong Language, and it's the creation of James Harbeck and Stan Carey.

James and Stan have enlisted a great lineup of contributors (I'm happy to be one of them). As the "About" page explains, Strong Language "gives a place for professional language geeks to talk about things they can’t talk about in more polite contexts. It’s a sweary blog about swearing."

Apologies to anyone (and it must be lots of you) who tried to reach a LLOG page yesterday and got redirected to x.vindicosuite.com.

This was the result of the latest malfunction in the sitemeter.com tool for counting visits and referrals, which we've been using for the past decade. Increasingly often over the past year or so, the sitemeter tracking code has been non-deterministically routing visitors to unwanted advertising sites, playing strange background music, etc. At about the same time, the company stopped responding to any support queries or complaints. Because its tracking statistics are useful, and because the unwanted redirections were rare and intermittent (and arguably due to mistakes rather than malice), I've stuck with them.

But as of yesterday evening, for a significant period of time, every single attempt to access a LLOG page resulted in a glimpse of the desired page followed quickly by redirection to x.vindicosuite.com, which is apparently some sort of passive DNS replicator or something. As far as I can tell, no virus or worm attack was involved, but the redirection alone is unacceptable, even if this is just another bug in sitemeter's counting software rather than anything malicious.

It seems that a lot of other people had the same problem with sitemeter (see also here, and many other comments over the past couple of years). So I've removed the sitemeter code from our WordPress installation. Now I can look forward to wasting a few hours trying to get sitemeter to stop charging me for their "service".

This is an interesting question raised by the Writing Chinese project at Leeds.  Helen Wang mentioned it to me in the hope that I might be willing to share my thoughts.  I'll do Helen one better and share this with many others, in hopes that they too may be willing to share their thoughts.