I can't tell if that is profound or trivial. I am basing it on the idea that recognition is a hierarchical process just as is classification Typical logical classifications are given by the subset relation, so it is a hierarchy. Eg: he is a football player, a human, a species, a living organism, etc.
The proposition in the title implies that each new branch point in the current classification hierarchy is associated to a specific "recognition measurement". Hence, as you make the sequence of sub-classifications bringing you to a final class, you will have made the same number of recognitions. Top down. Hence:
To know it is to store it
This was brought back to mind by watching a YouTube video from asapSCIENCE about how the Greeks could not see blue - or rather had no word for it. For the Greeks, blue was just another species of "dark" along with what we call "black". The video makes the point that it was in later languages that "blue" became separated from the more general "dark". This is, I believe, a point made by Worf - that you cannot really perceive blue as different from black if you do not have the vocabulary. It also implies some nasty realities about racism.
Also of interest to me is the long term question about how to build a recognition/classification hierarchy. I spoke about Sphinxmoth: forced completion and how the concept tried to describe the moment during learning when a single class in the classification is sub-divided. It seems to me this is what happened, after the Greeks, when a language did start using the word "blue".
To avoid being a pseudo intellectual, let me mention "Best Models" classification. What it says is that you have to factor out scale, then fit a (factor-parameterized) model to the given. This gets at the deeper relation between patterns and scale. That is missing from the above discussion. These "recognition measurements" are somehow orthogonal to scale. And I don't get it yet.