"A is like B" is a predicate of an argument. I'll give you an example of the application of logic in action on a set of predicates, which may help:
"A is like B" and "B is like C", therefore "A is like C".
Notice my capacity to do a formal logical sorite(s) here without having to understand anything about what "like" means.
You have done no such thing. "A is like B" is not in the form of a categorical proposition.* Like is not a logical operator. It is a shorthand for a series of logical operations.
A: Emmit Smith
B: OJ Simpson
C: Ted Kennedy
Emmit Smith is like OJ Simpson. OJ Simpson is like Ted Kennedy. Therefore, Emmit Smith is like Ted Kennedy.
It doesn't work. "is like" has different meanings in the two predicates. Logical operators don't change meaning from predicate to predicate. In the first, "is like" refers to the ability to rush a football. In the second, "is like" refers to the ability to get away with homicide without being convicted.
That is why "like" can be subsumed. It is not a logical operator, but a series of operations, and is subsumed by replacing the word with the series it represents. You are confusing an english word with a logical operator. Using your arguement, I could claim "as bad as" is a logical operator: A is as bad as B, B is as bad as C, therefore A is as bad as C. Not every word or phrase is a logical operator. But some words do represent chains of operations. It's easier to say "like" than to lay out the whole chain of reasoning behind it as long as speaker and hearer have the same chain of reasoning in mind. If not, then things have to be made explicit. We use "enthymic" words all the time. This in no way, shape, or form means that anything other than logic was behind the reasoning that went into the arguement.
*If you insist "A is like B" is in predicate form, your soritie must become: (A) is (like B), (like B) is (like C), therefore (A) is (like C). Now all you have to do is define the relationships between "B" and "like B" and between "C" and "like C". All this can be done logically. "like" is nothing more than a plain language shorthand so we don't have to have discussions like this all the time.
Than list them (in the order in which they can be applied to produce "like" as a theorem that can be so applied), please, without another ardent non-technical lecture that avoids this absurdly simple task--provided you are correct--yet again.
No, I'm not. What you are mish-mashedly referring to is the very problem of "domain of discourse" I have been referring to constantly. If you use different meanings of "like", you are referring to different domains of discourse and pretending logical operators apply across these domains.
Many ancient greek riddles and 12th century proofs of God's existence rely on this trick. It's true, I didn't say "tap, tap, no straying from the subject", but that is generally assumed by mathematicians in talking formally about sets.
Your argument, if examined in the light of day, should convince you that you are a little confused here. How can you claim "like" has several different meanings, and in the same breath, claim that "like" is just a shorthand for a series of logical operators?
When you use the world "like" in the sense of drawing an analogy, in fact, that is precisely it's logical weakness. What you are doing with the "like" operator is intentionally mixing domains of discourse, because intuitive insight may be a more important than preserving logical precision.