Language Log

On July 5, Phil Resnik delivered a keynote address at the 64th Annual Meeting of the Association for Computational Linguistics, with the title "A New Balancing Act: Reflections on the Relationship between Computational Linguistics and AI":

In this talk I argued that the field of computational linguistics – a term that includes NLP as its engineering-research subdiscipline – is experiencing a “success catastrophe”. The commercial success of LLM-based AI has thrown three key aspects of our research community out of balance. Here are three balancing acts we face:

First: Like any research community, we can recognize and take advantage of the knowledge obtained in earlier generations of work; we can also lean into new approaches.

Second: We can focus on language as language, which is to say, the properties of language that make it distinctive and human; we can also treat language as an input/output modality for AI systems.

Third: We can emphasize our role as a research community, where our primary purpose is to contribute to the stock of human knowledge; we can also emphasize our role in making sure that our young members have a path forward to get jobs – particularly jobs in industry since the path into academia is never a sure bet and for many of them industry is the goal.

In each of these pairings, the central importance of the former has given way to the overwhelming dominance of the latter.

Following up on our recent post about generational punctuation differences, here's a cartoon from the 7/27 New Yorker:

Below is a guest post by Richard Sproat and Rafael Núñez. It summarizes a recently published paper by d’Errico, Santos da Rosa, Courtenay,  Núñez,  Liu,  Sproat, & Blasi,  "Inferring Conventional Sign Systems from Paleolithic Engravings: Methodological and Theoretical Challenges".

Back in February, Christian Bentz and Ewa Dutkiewicz (henceforth BD) published a paper in the Proceedings of the National Academy of Sciences, that presented their study of markings on Paleolithic artifacts, mostly in bone, antler and ivory, from the Swabian Aurignacian (43,000–34,000 BP). Their conclusion, based on various statistical analyses, including entropic measures, and a couple of sign repetition rate measures, was that these marks constituted a structured system of deliberate and conventional signs. By translating the engravings into 1D sequences of standardized characters, they reported strong statistical overlap with early Mesopotamian protocuneiform (Uruk V) and clear divergence from later systems as well as modern writing systems.

Apparently so, though it's continuing to evolve. Reviewing Florence Hazrat's forthcoming book On the Mark: From Periods to Interrobangs, How Punctuation Remade the World, Judith Shulevitz focuses on the current generational and contextual differences in English punctuation ("Punctuation: A generational divide", The Atlantic 7/14/2026).

There was a flurry of attention to such things a couple of decades ago:

"The new semiotics of punctuation", 11/7/2012
"Agressive periods and the popularity of lingusitics", 11/26/2013
"Generational punctuation differences again", 8/1/2014
"Query: Punctuation in personal digital media", 2/23/2015

This question arrived recently by email:

A topic I wonder about: "annoying" words, and the cause of their annoyingness, and whether many people find the same particular words annoying, or if there's a long tail of very few or just one person who finds them annoying.

The broad category is "workplace jargon", but not in the "normal sense" of that ("brown bag lunch", "team player", etc. etc.)

Examples:
– I find the word "remit" (as in: what one is supposed to DO as a job function deliverable) quite, quite irritating. I don't know why. It sounds slightly stodgy, but I like other forms of stodginess.
– I find the word "mandate" equally annoying, when used in the same sense: that which one's job description requires them to do. It means the same thing! Is it because it seems to be a Britishism, while I speak American?

Is it that I dislike the things required of one's job? I really don't think that's it, but maybe my individual experiences with interactions happening upon these "remit" and "mandate" usages are the simple reason for the negative association?

Is there another reason beyond personal priors that I'd find "remit" and "mandate" out of place and annoying in the modern American workplace? Is there a category of words like this? It's more than a simple pet peeve, I think.

I did try web search and LLMing. Consulting an LLM, I find it's not "corporate BS", "pretension, or "weasel words" that bother me. Seems it's some kind of "legalese encroachment" that feels odd. I seem to not mind examples of military or medical encroachments to the same level. It's the authority that's common to them, the LLM says. Yet I don't take too much issue with: caveat, deconflict, fiduciary, jurisdiction, redact, sanction, arbitrate.

I'm going to call them "lightning-pole words", for lack of a better phrase.

Thursday evening, President Trump gave a speech about election security, and since it was only 25 minutes long, I listened to the whole thing rather than relying on summaries. I'll leave the political analysis to others — what struck me was a speech error that happened a little more than half way through, at about 14:49:

Tonight we're publishing a series of
previously classified U.S. intelligence community assessments
and other reports proving that our
government has wrong known these machines are
extremely exposed to attack.

Today's NonSequitur:

Bob Shackleton sent a link to this headline: "After 700 Years Underwater Divers Recovered 80-Ton Blocks from the Long-Lost Lighthouse of Alexandria", ZME Science 7/5/2026. His comment: "That’s a long time for a diver to stay underwater…"

There's an interesting video (tiktok or YouTube) from Alisa @multilingual_nest that starts with the question "Can you guess where I'm from based on what I'm about to tell you about my accent?" Here's Alisa's audio:

Anthropic's Claude gives good programming help, in my experience. And it can offer good explanations of obscure phrases and concepts. But its apparent assumptions about (implicit) conversational common ground are often extremely odd, as the following example illustrates.

Yesterday's news was full of stories about how the U.S. Government had agreed to let Anthropic release Fable 5. However, my Claude app still opened conversations with a note telling me that Fable 5 was not available. So I asked when that would change, and got a weird answer (from the Sonnet 5 version):

Victor Mair passed away two days ago, on Sunday 6/28/2026. We'll have a proper series of memorial posts and links later, but for now, here's a Facebook post from his brother Denis, his Wikipedia page, and a link to LLOG posts written by or mentioning him.

Over the past week, there have been several thousand new readers of a series of posts from five years ago, apparently because they were referenced in a popular place.

The last six panels of Yesterday's Doonesbury: