Try it now: Tap your hand on the desk randomly. Can you recall how many times you did it without "saying" a sequence in your head like "1, 2, 3" or "A, B, C" etc?
If yes, how far can you count? With a language it's effectively infinite. You could theoretically go up to "1 million 5 hundred 43 thousand, 2 hundred and 10" and effortlessly know what comes next.
I spent the next few days trying to understand how that process worked. I would force myself to think in words and sentences. It was incredibly limiting! So slow and lacking in images, in abstract relationships between ideas and sensations.
It took me another few years to realise that many people actually depend on those structures in order to produce any thought and idea.
We recently put the project I've been working on for the last year out into the field for the first time. As was fully expected, some bugs emerged. I needed to solve one of them. I designed a system in my head for spawning off child processes based on the parent process to do certain distinct types of work in a way that gives us access to OS process-level controls over the work, and then got about halfway through implementing it. Little to none of this design involved "words". I can't even say it involved much "visualization" either, except maybe in a very loose sense. It's hard to describe in words how I didn't use words but I've been programming for long enough that I pretty much just directly work in system-architecture space for such designs, especially relatively small ones like that that are just a couple day's work.
Things like pattern language advocates aren't wrong that it can still be useful to put such things into words, especially for communication purposes, but I know through direct personal experience that words are not a necessary component of even quite complicated thought.
"Subjective experience reports are always tricky, jerf. How do you know that you aren't fooling yourself about not using words?" A good and reasonable question, to which my answer is, I don't even have words for the sort of design I was doing. Some, from the aforementioned pattern languages, yes, but not in general. So I don't think I was just fooling myself on the grounds that even if I tried to serialize what I did directly into English, a transliteration rather than a translation, I don't think I could. I don't have one.
I'm also not claiming to be special. I don't know the percentages but I'm sure many people do this too.
An example of this would be when I’m lifting weights with a friend and am lost in the set/focusing on mind-muscle connection, and as a result I forget to count my reps. I am usually quite accurate when I verify with my lifting partner the number of reps done/remaining.
As OP mentioned, many people have no internal speech, otherwise known as anendophasia, yet can still do everything anyone with an internal dialogue can do.
Similarly for me, I can do “mental object rotation” tasks even though I have aphantasia.
For context I have both abstract "multimedia" thought processes and hypervisor-like internal narrative depending on the nature of the experience or task.
X . . X . . X . . . X . X . . .
and every so often switch out for variations, eg: X . . X . . X . X . . . X . . .
or X . . . X . . . . . X . X . . .
but I'm no good for playing polyrhythms, which many other people can do, and I believe they must also do so more by feel than by counting.If I want to translate this knowledge into a number, I need to count the taps I am seeing in my head. At that point I do need to think of the word for the number.
I could even do computations on these items in my mind, imagine dividing them into two groups for instance, without ever having to link them to words until I am ready to do something with the result, such as write down the number of items in each group.
https://www.sciencealert.com/theres-a-big-difference-in-how-...
X . X X X . X . X X X .
A . . A . . A . . A . .
B . B . B . B . B . B .
and: X . . X . X X X . X X . X . X X . . X . X X . . X X . X X . X . . X . X X . . X X . X . . X . . X X X X . . X X X X . . X . . X . X X . . X X . X . . X . X X . X X . . X X . X . . X X . X . X X . X X X . X . .
A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . . A . . . .
B . . . . . . B . . . . . . B . . . . . . B . . . . . . B . . . . . . B . . . . . . B . . . . . . B . . . . . . B . . . . . . B . . . . . . B . . . . . . B . . . . . . B . . . . . . B . . . . . . B . . . . . .
C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . . C . .
Learn to do them with one limb (or finger) per line, and also with all the lines on the same limb (or finger). And then suddenly, they'll start to feel intuitive, and you'll be able to do them by feel. (It's a bit like scales.)Other animals with at best very limited language, are still highly intelligent and capable of reasoning - apes, dogs, rats, crows, ...
...maybe I do this sometimes myself. Remembering the proper names of things is effort.
It's the equivalent of <thinking> tags for LLM output.
I would note though I have a really difficult time with arithmetic and mechanical tasks like counting. Mostly I just lose attention. Perhaps an inner voice would help if it became something that kept a continuity of thought.
This is a parallel stream, because if I count with imagined pictures, then I can speak and listen to someone talking without it disturbing the process. If I do it with subvocalization, then doing other speech/language related things would disturb the counting.
Also, many people simply repeat facts they were told. "We need words to think" is simply a phrase this person learned, a fact to recite in school settings. It doesn't mean they deeply reflected on this statement or compared it with their experience.