Language Log

According to Wikipedia, word salad

is a "confused or unintelligible mixture of seemingly random words and phrases", most often used to describe a symptom of a neurological or mental disorder. The name schizophasia is used in particular to describe the confused language that may be evident in schizophrenia. The words may or may not be grammatically correct, but they are semantically confused to the point that the listener cannot extract any meaning from them. The term is often used in psychiatry as well as in theoretical linguistics to describe a type of grammatical acceptability judgement by native speakers, and in computer programming to describe textual randomization.

The phrase {word salad} has become increasingly common recently in the popular press, most often as an insulting description of Donald Trump's spontaneous speech. See for example Sahil Kapur and Peter Nicholas, "'Incoherent word salad': Trump stumbles when asked how he'd tackle child care", NBC News 9/6/2024.

"The mysterious slowdown of VPNs in China", Drum Tower, The Economist (newsletter), Gabriel Crossley (China correspondent) (8/15/24)

Every summer Communist Party bigwigs leave Beijing and go to recharge in the resort town of Beidaihe on the coast. This coincides with the silly season for China watchers. There is little hard news, so rumours fly. Some are baseless speculations about the health of Xi Jinping, China’s supreme leader. Others are more well-founded, such as a report that Hu Xijin, a prominent nationalist commentator, has been muzzled on social media (probably for accidentally overstepping the party line). And some are true but harder to explain: like the fact that virtual private networks (VPNs) are getting slower.

It seems that ChatGPT still has a few things to learn, about conversational dynamics as well as about interlocutor modeling:

@risthinks ChatGPT chatting each other about AI ! #AI #ArtificialIntelligence #ChatGPT #TechTalk #FutureTech #Conversations #Innovation ♬ original sound – RisThinks

From "Lost Futures: A 19th-Century Vision of the Year 2000", The Public Domain Review, 6/30/2012:

What did the year 2000 look like in 1900? Originally commissioned by Armand Gervais, a French toy manufacturer in Lyon, for the 1900 World exhibition in Paris, the first fifty of these paper cards were produced by Jean-Marc Côté, designed to be enclosed in cigarette boxes and, later, sent as postcards. All in all, at least seventy-eight cards were made by Côté and other artists, although the exact number is not known, and some may still remain undiscovered. Each tries to imagine what it would be like to live in the then-distant year of 2000.

Language Is Not Enough for Brains in Conversation

Zoom Webinar: https://uu-se.zoom.us/j/69177119780

4 October, 2:15 p.m. SEMINAR – WEB EVENT

Swedish Collegium for Advanced Study (SCAS), Uppsala

Julia Uddén, Pro Futura Scientia Fellow, SCAS, and the Departments of Linguistics and Psychology, Stockholm University.
Affiliated Researcher, Department of Neurobiology of Language, Max Planck Institute for Psycholinguistics, Nijmegen

Riding the trolley from West Philadelphia going to University City,

https://twitter.com/slurve/status/563688361316339713

A key term in Chinese historical geography is guān 關 ("pass").  You can see from the shape of the character that it is framed by the two panels of a door, left and right, and that it has two upright, elaborated bars that could impede progress through the gate (I am thinking of the early forms of the character).  The flanking door panels constitute the semantophore (radical, classifier) of the character, and the bars inside are the secondary semantophore, but may also simultaneously function as a phonophore.

A pass serves both to facilitate and block movement along key routes leading into and out of a country or regions within a country.

Just as I was thinking about writing this post on passes, I synchronously and serendipitously received from Alan Kennedy a reference to this highly technical article on Silk Road travel:

Irina Tupikova, Matthias Schemmel, Klaus Geus, "Travelling along the Silk Road: A new interpretation of Ptolemy’s coordinates", Max-Planck-Institut für Wissenschaftsgeschichte / Max Planck Institute for the History of Science, Preprint 465 (2014), 73 pages.

A month ago ("Real people in virtual worlds: a viral update?", 3/5/2020), I noted that

[T]he popular virtual-meeting applications don't yet have a way for a group to hold their discussion in a shared virtual space, as in current video games or applications like vrchat.

And when participants' avatars (realistic or otherwise) can sit or move in a shared space, with appropriate directional audio and so on, we'll be able to have virtual seminars, virtual workshops, virtual corridor conversations — and most important, virtual dinner parties!

One positive outcome of the growing panic over COVID-19 will be to hasten the deployment of these technologies.

So today I learned that some University off Pennsylvania students are creating a virtual Penn campus, with the idea of holding (a version of) events there like Hey Day and the Penn Relays — "UPenn students recreated their campus on 'Minecraft' in painstaking detail while stuck at home — take a look", Business Insider 4/5/2020.

Look everyone! it's a post about language in China by not-Victor! :)

I just had to drop everything and write this post while I was listening to the latest Reply-All podcast, this week consisting of a series of phone interviews with people around the world about the experience of the COVID-19 pandemic in their area. The first interview was with Justine from Shanghai, and she was talking about ways people were working around censorship in talking about…

…uh oh I suddenly realize I may be doing a disservice to the Chinese public by posting about this, so I won't go into all the detail I intended. Anyway, the basic idea is that folks were using homophonic transliteration with emojis to get around censorship of certain stories about the epidemic there. You can listen to the podcast here; the relevant bit is between minutes 4:40 and 6:15.

If you can imagine it, this would be like trying to parse "Little Red Riding Hood" from emojis  like💡💯🐀✍️👱‍♂️. Leaving comments open to see if anyone can figure out what homophonic transliteration words I intend for that sequence. First prize is a disinfected plastic cup with logo from the Language Log water cooler stash delivered by drone sometime in 2022.

It's probably worth noting that this idea of communicating via pictures of sound-alikes is basically the actual honest to god origin of phonetically based writing systems. Also worth noting that this way of repurposing symbols to represent sounds of another expression has a long history particularly in Chinese and related languages, whose linguistic features mean that you often have lots of homophones and near-homophones, and whose logographic writing systems probably lend themselves to that kind of graphemic/phonemic cross-indexing during lexical lookup. (Someone must be studying that, right? ) So you get a lot of punning and double-entendres in Chinese writing, if I understand rightly.

If the idea of homophonic transliteration is new to you,  you could get in the wayback machine and check out this archival LL post from simpler times, here.

Wishing all my fellow humans the very best from a living room in Arizona!

This period of prudent isolation is a good time to remember that linguistic analysis applies not only to sound, structure, and sense, but also to social interaction. As the first in a series of posts on this topic, we feature Eve Armstrong's brilliant application of simulated annealing to a problem currently on hold, but sure to re-emerge in full force when our lives de-virtualize: "An Artificially-intelligent Means to Escape Discreetly from the Departmental Holiday Party; guide for the socially awkward" (4/1/2020)G:

We shall employ simulated annealing to identify the global solution of a dynamical model, to make a favorable impression upon colleagues at the departmental holiday party and then exit undetected as soon as possible. The procedure, “Gradual Freeze-out of an Optimal Estimation via Optimization of Parameter Quantification” – GFOOEOPQ, is designed for the socially awkward. The socially awkward among us possess little instinct for pulling off such a maneuver, and may benefit from a machine that can learn to do it for us.

In the past few days, I've encountered some nice examples of the communicative interpretation of what I've suggested we ought to call "interpolations" rather than "disfluencies".

Article in The Washington Post (6/18/19):

"Robocalls are overwhelming hospitals and patients, threatening a new kind of health crisis"

Last week ("Joos jokes", 8/14/2018) I linked to the "Proceedings of the Speech Communication Conference at M.I.T.", published in 1950 in the Journal of the Acoustical Society of America,  and I promised to revisit "this window into a bygone age" in a later post. So today I present to you the following passage from "Introduction: A Definition of Communication" by S. S. Stevens:

Although no phenomenon is more familiar to us than communication, the fact of the matter is that this magic word means many things to many people. A definition broad enough to encompass all these meanings may risk finding itself dissipated in generalities, but for the purposes of this conference a broad operational definition of communication is, I believe, both appropriate and possible. I should like, therefore, to venture the following: Communication is the discriminatory response  of an organism to a stimulus.

He goes on to confirm that this perspective is indeed as weird as it seems:

This definition says that communication occurs when some environmental disturbance {the stimulus) impinges on an organism — and the organism does something about it (makes a discriminatory response). If the stimulus is ignored by the organism, there has been no communication. The test is differential reaction of some sort. The message that gets no response is not a communication.

This definition is broad, operational, and behavioristic. It includes the anxious clucking of the mother hen, which brings the chicks scurrying to shelter. At a different extreme it includes the modem treatise on information theory, which some people seem to read and respond to with a glow of understanding. By appealing to behavioral operations as the test for the presence or absence of communication, we explicitly forsake all concern with abstracted meanings, significations, and the like, unless, of course, these words are in turn defined in terms of discriminatory responses. In short, we stick to observable phenomena.

Are we really supposed to believe that some people's "glow of understanding" on reading about information theory is an "observable phenomenon", sufficient to characterize what has been communicated? This strikes me as a reductio ad absurdum of Stevens' behavioristic prejudices — as effective an argument as Noam Chomsky's 1959 Review of B.F. Skinner's Verbal Behavior.