semantic tensors

A tensor is a device which physicists and mathematicians use when they want to represent a quantity or bunch of quantities which retain their values or some other kind of invariant under a transformation of coordinates. Therefore, a semantic tensor is a device that semantic engineers or the Jane on the street use when they want to represent a meme in an interpretation-invariant way. An important thing to consider about semantic tensors is that they are related to, but different from, encryption. Encryption, on the whole, doesn't take into account those who use it: that is to say, that encryption requires a knowledgeable person applying the decryption, or an unintelligent person with just enough understanding of the encryption algorithm. In any case, the two are bound by a certain amount of having-one's-head-screwed-on-properly. But it isn't possible to make a normative gloss of this: which is to say that there's not one canonical way to agree on who has their head screwed on correctly: in a world with diverse amount of intellectual activity there are going to be a wide variety of people, and as a result, brainy people will entertain a variety of opinions. You can either be explicit about it, like Mensa, or implicit about it, like any number of putative secret societies. In any case, this leaves you with a mess. The way out of this mess is to decide to construct objects which are complicated, and whose complexity is addressible by perceptases: you make an object or thing that is gnarled, and make that gnarl the type of gnarl that you're interested in, realizing that the object is a product (in the category theoretic kind of sense) of variously interpenetrating types of gnarl, and that those whose perception is geared for that kind of gnarl, it will scream out "pay attention to this object" or "this is signal, decode it", and so on and so forth. Ensconced in puzzles, this is the thing which I intend to do. I will make multiply orthogonal high gnarl objects in such a way that each piece of high gnarl decodes to the same thing, and that is how I will create semantic tensors.