I had a hypothesis about my use of “internal speech” and used that game to give myself a nominal task to see if I heard any internal speech while engaging with it (and if so with what content). Prior to playing with the game, I had a sense that I rarely use elaborate and intelligible “internal speech” except in the context of high impact communication, as when rehearsing a speech for an audience, planning to write prose, retrospectively mulling over an emotionally fraught conversation, or trying to figure out what I’m capable of saying in a second language. In those cases I imagine saying/writing something to see how it sounds, and then imagine alternative things I could say instead.
For normal daily life, as near as I can tell, I simply think in images, kinesthetic feelings, and gestalt model/analogy/diagram/datastructures. I didn’t come to this by any strategic plan that I know of, I’ve just always had these mental tendencies. My earliest memories (that are clearly not photographs I’ve re-written as episodic memories) are not words and stories, but are about (or associated with) floorplans and the texture of carpets and furniture.
The idea of “hearing all your thought” seems weird enough to me that I wonder if its a straw man? Or perhaps this is one of those places where people tend to generalize from one example more than they should? I would not naively expect people to be the same in this respect… For reference, I’m mediocre at math, but very verbal according in aptitude testing contexts. I maxed the SAT verbal, participated in (and later coached) policy debate, placed very high in two spelling bees when I was a child, and don’t mind playing with morphological rules to make up words that should be in the dictionary, even if they are not. I can juggle reasonably well and am not any worse at social processes than I want to be, given my ethical tendencies.
My usual lack of internal speech doesn’t happen because I’m bad with words in general, it is just that I don’t find them to be a particularly expressive medium for transmitting most of what is in my head. When I was growing up my family kept piles of scratch paper and pencils in a standard place in the kitchen and living room so that if a conversation became substantive we could sketch our thoughts and use pointing and words and images to communicate what was in our heads. For example, a conversation about gardening might naturally turn into sketching diagrams of vegetable layouts, with 3-D projections and arrows to show how the sun was expected to fall based on nearby trees, and a second diagram to show how the garden would evolve after 3 months as plants grew and died. Those details are hard to discus but easy to diagram.
When playing the game you linked to, I found that sometimes I’d use internal words like “blue red” while I was looking at (for example) the left side of a hole and the bottom side the hole, so that I could “hold the sounds in my audio register” while looking at tiles to see if a particular tile (based on its left and bottom colors) would fit in the hole I had looked at earlier. When I found a matching tile I didn’t think “Now I put this in the bottommost opening on the right”, instead I’d just move it to the place my kinesthetic memory said that I had previously seen the hole I was trying to fill.
I would bet that the trick of using an audio register to run working memory intensive mental algorithms is probably very very common. Somewhere (unfortunately I forget the cite) I read that that in WWII the near universality of using words for doing math in the language you learned arithmetic was used to suss out spies. Like people who had perfect English/American accents would be asked to talk out long division, and the German spies would fall into muttering “ein, zwei, drei...” and thereby reveal themselves to have been educated in a German speaking school system.
In the tile game, other than using my audio buffer for extra working memory, I found myself either noticing that I had silent stretches where I wasn’t making internal speech, but it also seemed that my attention to internal speech was causing me to produce it. Mostly I was trying out phrases to use in this comment to describe what happened in parrallel with actually playing the game. A few times I noticed myself internally commenting on my performance (eg “that was dumb”, or “how many bits to uniquely specify a tile” with a rough calculation of four colors x two edges = roughly enough to code for all the tiles) in ways that would lead to optimizations in my playing behavior (by cutting out behavioral tendencies that I could articulately criticize or engaging in optimizations I could verbally suggest). So internal speech wasn’t entirely absent, it just wasn’t an internal monologue taking the form of “my whole internal mental tableau”. It happened mostly in parallel with game performance, rather than narrating the performance.
Now that I think about it, I remember about two years ago I played a non-standard memory game (flipping facedown cards face up two at a time and keeping them iff they are have the same image) and I got a pretty large performance boost from turning them into flashcards for 30 minutes where I counted them as memorized if I could look at the image and connect it to a unique summary phrase like “mohawk kitty” or “fat chair cat” or whatever. The problem is, the language itself might not have been critical there? Perhaps it was just enough image familiarization that visual chunking based on memorized subtle features was easier? I bet an experiment could be designed to test this stuff...
Try playing this game.
Excellent testbed!
I had a hypothesis about my use of “internal speech” and used that game to give myself a nominal task to see if I heard any internal speech while engaging with it (and if so with what content). Prior to playing with the game, I had a sense that I rarely use elaborate and intelligible “internal speech” except in the context of high impact communication, as when rehearsing a speech for an audience, planning to write prose, retrospectively mulling over an emotionally fraught conversation, or trying to figure out what I’m capable of saying in a second language. In those cases I imagine saying/writing something to see how it sounds, and then imagine alternative things I could say instead.
For normal daily life, as near as I can tell, I simply think in images, kinesthetic feelings, and gestalt model/analogy/diagram/datastructures. I didn’t come to this by any strategic plan that I know of, I’ve just always had these mental tendencies. My earliest memories (that are clearly not photographs I’ve re-written as episodic memories) are not words and stories, but are about (or associated with) floorplans and the texture of carpets and furniture.
The idea of “hearing all your thought” seems weird enough to me that I wonder if its a straw man? Or perhaps this is one of those places where people tend to generalize from one example more than they should? I would not naively expect people to be the same in this respect… For reference, I’m mediocre at math, but very verbal according in aptitude testing contexts. I maxed the SAT verbal, participated in (and later coached) policy debate, placed very high in two spelling bees when I was a child, and don’t mind playing with morphological rules to make up words that should be in the dictionary, even if they are not. I can juggle reasonably well and am not any worse at social processes than I want to be, given my ethical tendencies.
My usual lack of internal speech doesn’t happen because I’m bad with words in general, it is just that I don’t find them to be a particularly expressive medium for transmitting most of what is in my head. When I was growing up my family kept piles of scratch paper and pencils in a standard place in the kitchen and living room so that if a conversation became substantive we could sketch our thoughts and use pointing and words and images to communicate what was in our heads. For example, a conversation about gardening might naturally turn into sketching diagrams of vegetable layouts, with 3-D projections and arrows to show how the sun was expected to fall based on nearby trees, and a second diagram to show how the garden would evolve after 3 months as plants grew and died. Those details are hard to discus but easy to diagram.
When playing the game you linked to, I found that sometimes I’d use internal words like “blue red” while I was looking at (for example) the left side of a hole and the bottom side the hole, so that I could “hold the sounds in my audio register” while looking at tiles to see if a particular tile (based on its left and bottom colors) would fit in the hole I had looked at earlier. When I found a matching tile I didn’t think “Now I put this in the bottommost opening on the right”, instead I’d just move it to the place my kinesthetic memory said that I had previously seen the hole I was trying to fill.
I would bet that the trick of using an audio register to run working memory intensive mental algorithms is probably very very common. Somewhere (unfortunately I forget the cite) I read that that in WWII the near universality of using words for doing math in the language you learned arithmetic was used to suss out spies. Like people who had perfect English/American accents would be asked to talk out long division, and the German spies would fall into muttering “ein, zwei, drei...” and thereby reveal themselves to have been educated in a German speaking school system.
In the tile game, other than using my audio buffer for extra working memory, I found myself either noticing that I had silent stretches where I wasn’t making internal speech, but it also seemed that my attention to internal speech was causing me to produce it. Mostly I was trying out phrases to use in this comment to describe what happened in parrallel with actually playing the game. A few times I noticed myself internally commenting on my performance (eg “that was dumb”, or “how many bits to uniquely specify a tile” with a rough calculation of four colors x two edges = roughly enough to code for all the tiles) in ways that would lead to optimizations in my playing behavior (by cutting out behavioral tendencies that I could articulately criticize or engaging in optimizations I could verbally suggest). So internal speech wasn’t entirely absent, it just wasn’t an internal monologue taking the form of “my whole internal mental tableau”. It happened mostly in parallel with game performance, rather than narrating the performance.
Now that I think about it, I remember about two years ago I played a non-standard memory game (flipping facedown cards face up two at a time and keeping them iff they are have the same image) and I got a pretty large performance boost from turning them into flashcards for 30 minutes where I counted them as memorized if I could look at the image and connect it to a unique summary phrase like “mohawk kitty” or “fat chair cat” or whatever. The problem is, the language itself might not have been critical there? Perhaps it was just enough image familiarization that visual chunking based on memorized subtle features was easier? I bet an experiment could be designed to test this stuff...
Anyway, hope my data point is helpful :-)