Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
What Is the Sound of Thought? (mitpress.mit.edu)
68 points by the-enemy on Sept 22, 2020 | hide | past | favorite | 44 comments


> Remarkably, we found that the shape of the electric waves recorded in a non-acoustic area of the brain when linguistic expressions are being read silently preserves the same structure as those of the mechanical sound waves of air that would have been produced if those words had actually been uttered

the Rosetta stone idea to map thought from languages waves (sound and light) is an interesting field of research

I have no idea how but since I was young I learned that I can read faster when I don't "voice" out the sound of thought

I also learned that as a multilingual person the "language" of choice for my thought can be different at times, even though I have no actual fondness of using any specifically


For me, some words and phrases sound better in one language than the other. So in my internal monologue, I often flip back and forth between Polish and English, sometimes multiple times in a single sentence. It feels like "the path of least resistance".

I also do that sometimes when talking, when I know the other person is proficient in English. Not everyone appreciates this, but some of the best conversations in my life were with a particular person who was like me in this regard; we'd just stream our thoughts to each other in this weird mix of Polish and English, that was hard to follow for anyone else.


I know the same feeling with English and German. It really is a wonderful thing. I think this would be called Code-switching [0].

> In linguistics, code-switching or language alternation occurs when a speaker alternates between two or more languages, or language varieties, in the context of a single conversation.

[0]: https://en.wikipedia.org/wiki/Code-switching


Wechseln even dans a fraza: http://www.europanto.be/cabillot1.html


Haha.

Minä gör the same a menudo.

Yo teen det samma often.

I hago the same ofta.

Jag do lo mismo usein.

(FI/SV/EN/ES) :)


Belta loda! You sound like your speaking the Belter slang from the scifi series The Expanse. Set a few hundred yrs into the future, slang used by the working class out mining the asteroid belts.


She she taki taki beltawala. Mi finyish vedi da lang belta da diye de.

https://www.youtube.com/watch?v=hCyGIJl1XUA


Taki taki. TIL. Tolowda wanya pochuye fo walowda sowngit ke? https://www.youtube.com/watch?v=_ULICEPpoLQ


Wow, closest I can get to that mix is: https://www.youtube.com/watch?v=6VqUiSdd3yM , an english band interpreting a finnish tango.

Swedish will have to be separate: https://www.youtube.com/watch?v=5hdUKZLQIuQ

(Are you a finland-swede or sweden finn? If either of those, I'm a fan of both Army of Lovers and Nightwish.)


Haha yes that Zappa gig is a classic!

"Swedish-speaking finn" or "Finland-swede" yes.


Y’all, stop. You’re confusing the hell out of GPT-4.


I hope they won't be training it on HN corpus, because I don't want an algorithm to spit out my own rants back at me...


Mi na du wanya wa robot fo du wating mi enjoy


It really is wonderful! Good to know there's an established term for that, thanks!


Now try to imagine instead of English, it's Russian? Switching from Polish to Russian. Will it be the same?


Polish/ukrainian song lyrics: https://www.youtube.com/watch?v=7cxciyZEBkE (switching between stanzas, not phrases, however)

Amateur joint translation here: https://news.ycombinator.com/item?id=24101696

(any native speakers care to say how close we came?)

(and notdang, on the off chance you're from somewhere around Moldova, what would you say the ethnicity of the girl in https://news.ycombinator.com/item?id=24406864 might be? I think a woman from Moldova is the normal reading, but apparently there's also an Odessa neighbourhood called Moldavanka...)


I know a little bit of Russian so a random word or two sometimes sneak into my speech or thinking process. I imagine if I were fluent in Russian, I'd switch back and forth just as well. They're more similar to each other than Polish and English, but not that similar as to be bucketed as one language.

I've been thinking about this phenomenon a lot in the past and I've noticed that two main triggers are a) emotional payload of a phrase, and b) subtle differences in meaning of words. Some phrases sound better to me in English than in Polish, and it's probably driven by the context in which I've learned them. As for words, it's often the case that even if a word in Polish corresponds directly to a word in English by dictionary, their etymologies and general use patterns draw slightly different borders in thingspace. Sometimes that distinction is important, and I automatically pick the language in which the word is closer to what I mean.


Not sure what are the sentiments in Poland, but in a couple of ex-USSR countries mixing Russian and the local language is regarded as a very bad taste. This applies exclusively to Russian.


> I have no idea how but since I was young I learned that I can read faster when I don't "voice" out the sound of thought

https://en.wikipedia.org/wiki/Speed_reading


The obvious follow up experiment is to have the subjects read silently first, then read aloud.

(I've also found I can readily manipulate algebraic symbols without having any idea of how they should be pronounced. splodge?)

https://i.pinimg.com/originals/67/bc/35/67bc35746db0758b5ab4...


>I have no idea how but since I was young I learned that I can read faster when I don't "voice" out the sound of thought

Subvocalisation is the term for that voice in your head when you read. I've never been able to get rid of it, and trying to has always resulted in me feeling like I'm skimming rather than reading.


I think a way to practice it would be through math exercises. You don't really voice out '3 + 8' as much as 'three plus eight'. Also thinking further I think books directed at children would probably also help, I feel like complex terms keep me subvocalising when reading.


yes, that is the secret.

"reading" is not one activity, there are 100s of ways to read.


I wonder if the same is true for languages where the writing system is not a representation of sounds, like written chinese.


My understanding was that written Chinese has a great deal of phonetic information (even though it is not exclusively phonetic like the alphabet or Japanese Hiragana/Katakana). I had understood that many ideograms essentially could be interpreted as topic + pronunciation: e.g. 驰 could be interpreted as "the word about 马 horses read as 也 chí".


At least one of the Chinese languages is tonal too, so the same sequence of vowel and consonant sounds can have totally different meanings when spoken at different pitches. (The humorous example is “grass mud horse,” which is a bit NSFW and a great act of someone thumbing their nose at censorship.)

I wonder if native speakers have a greater emphasis on subvocalized pitch, or if that’s even a part of subvocalizing? Paying attention to my own internal reading now, I suspect that it is.


English is my first language, but I find that when I read in Mandarin, I can focus on intonations, personality, etc or leave it out as sort of monotone.

The intonations that give the characters their meaning are so engrained in the flow of the sentences it can feel monotone, as it is just the way it is said.

If I am reading a poem or idioms I usually like to over emphasize the tone like a classic theatrical reading because it’s fun and helps me remember what I read.


I assume the reason for this is that speech is encoded as an internal meaning and the visible text is also encoded as meaning. When the circuit for the internal meaning fires up then the rest of the circuits are sympathetically fired as well. It also makes me assume that's why PTSD is a thing - some trigger hits the specific circuit that causes the rest of the stress response. With PTSD the process is debilitating, but in the normal course of life the process makes retrieval of information and the expression of appropriate actions easier.

Neat. I'm not a scientist and I could be very wrong however.


Show me data, the signals. Except for maybe short pauses between phonological clusters, I need extraordinary proof to accept such extraordinary claims.


I attach some references that could be related:

https://research.iusspavia.it/handle/20.500.12076/265

Or, despite being too recent to be in the book:

https://research.iusspavia.it/handle/20.500.12076/1463


This study assumes that everybody hear the sounds of words while thinking, and that's not the case.

Many people (although I don't know the statistics) do not have this "inner speech".


The observable result "the subject reports a lack of inner speech" can correspond to "the subject lacks inner speech" or "the subject isn't aware of inner speech" or "the subject have a motivation to not report having inner speech" or "the subject can't report having inner speech" or "the subject doesn't recognize his/her subjective experiences as inner speech". And maybe more.


I get that you are in disbelief, but this is actually a thing.

Numerous testimonies are unambiguous, such as this one:

https://www.youtube.com/watch?v=u69YSh-cFXY

Like, the interviewee is able to read extremely fast ("superhuman", according to the interviewer), as a consequence to the absence of inner speech.

This kind of details eliminate the possibility of the uncertainties you suggested.


From the article:

> Yet we know that language can also be present in the absence of sound, when we read... or when we use words while thinking — in technical terms, when we engage in endophasic activity.

> This simple fact immediately raises the following crucial question: What happens to the electric waves in our brain when we generate a linguistic expression without emitting any sound?

"Endophasic activity" is our "inner speech". A charitable interpretation of Andrea Moro's 16 person experiment is that they only included people with typical hearing. Experiments that target the atypical hearing cases may be crucial in understanding how the mind processes language. Indeed, the article goes on to state:

> Among the questions this discovery raises: What kind of electrical activity is elaborated in a language network (one that includes the Broca’s area) by persons who have never been able to hear any sound from birth?


I think you misinterpreted the use of "endophasic activity" here, it just means that the speech is not real, that it is inner, or the brain not producing any real efferent output, just internal-bounded output.

So, the study considers what happens when using linguistic thought without uttering words (aka inner speech), but it does not investigate the properties of the mind of people who do not experience inner speech.

For the second quote, it is maybe closer to the subject I raised, but there can be an homologous of "inner speech" for mute or deaf people (e.g. visualizing linguistic cues), so there is still the possibility of a category of people with and another without these cues.

Afaik, all these questions have been ignored by the study.

Not that it's bad, I mean, they have to restrict to some narrower question at first if they want to publish without spending decades on it.


It was the first time that I encountered the word "endophasic" so I looked it up before my first reply. It isn't included in my local dictionary but the first Google search result defines it as:

> Being or pertaining to endophasia (internal, unvocalized speech).

I only mentioned it so people didn't have to look it up like I had to; I don't think our interpretation differs.

I quoted the first couple of sentences from the last three paragraphs of the article. Those paragraphs emphasize the exciting questions raised by this small study, including the important question you raised.

To be fair I only read the article and I didn't search for the study but I look forward to reading the book:

> This article is adapted from Andrea Moro’s book “Impossible Languages.”


I'm curious if anyone has studied if brain activity during "awake surgery" is different in some major ways from normal brain activity (I'm not sure how you would measure normal brain activity to a comparable degree though).

It seems like "awake surgery" is an often cited tool for understanding how the brain works, but it is a pretty extreme situation, so I could imagine for instance, maybe the stress of awake surgery has a tendency to make the mind suppress certain types of thought which might otherwise be prominent.

Not saying this is the case, but if we are going to rely on the conclusions of awake surgery for important things, we should double check the method to make sure we are not overlooking anything important.


When I'm in the zone, thought sounds like a jet turbine "ffffffffffff", or being underwater. Sight stops working, or at least I don't remember what I saw during these moments. Also hearing seems to stop working, at least for outside sounds.

If I'm facing a screen, it is practically I see and I hear nothing.

If I get interrupted when I'm in that kind of trance, as in someone getting in my view starting to make signs, or poking me, I need a few moments to regain the ability to speak.


I have the same. But in my case it lasts for hours after work. Not being able to comprehend the speech of my kids. For me i got some label to put on it after being diagnozed with Skizotypia


If the subject's thoughts are then converted to audio and played back to the subject I wonder what fixed points would emerge.


Its same as the dreams are colorful or black and white


Fail


The article contains no examples of the "sound of thought". Epic fail.


I don't understand the downvoting; the MIT Press release could at least point to the research backing the claims, which I guess could be:

https://research.iusspavia.it/handle/20.500.12076/265

Or, despite being too recent to be in the book:

https://research.iusspavia.it/handle/20.500.12076/1463

(of the latter, I guess that this link: http://150.162.46.34:8080/embc-2015/papers/13891714.pdf is a summary)




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: