"Brevity is the soul of wit" suggests that the transaction granularity should be as high as possible. Also, for also for something to be funnier, it must be made understandable by a smaller number of people. The use of MIAOs decreases transaction granularity. Also, if jokes are made about a group of people, then stereotypes about Jews or hispanics or afro-americans form a MIAOplex that can decrease the transaction granularity of jokes about other races, that is, racist jokes or comments. The comments or jokes are easily made, even with little thought to "writing" them, hence even stupid people, memetically unsophisticated people, can make them. This is interesting since that provides a gateway to group membership for nominal outsiders, the stupid, the unsophisticated. It's easier for a stupid person to enter a racism enclave than it is, say, for a camel to enter a keyhole eye of a needle or for a rich man to enter the Kingdom of Heaven.
Okay, so what can we do about racism? Ultimately transaction granularity is linked to contact memes. Is this universally so? Well, in a racism paradigm, there are contact memes with respect which act as interface memes for the to the deprecated race from the main race. These contact memes allow for decreased transaction granularity for interaction between the two sub-cohorts.
Why the majority race, the ruling race, hate it when the subject deprecated races try to control their own narrative? Their own memetic system or rather to connect their own memetic system to the ruling race's system?
What does that look like if I get to come in and update your memetic system to include or accomodate my system (as a member of a minority), whether I'm a native american, transexual, gay, a woman (vis a vis male hierarchy), or whatever?
What does this look like, memetically?
Integration is difficult. Or it can be or it at least presents certain challenges. Potentially every memetic interaction is subject to redaction. When you're going to say "sir or madam" you now must say "xlehck" except in cases where multiple genders or skeezaks are present....
This requires a lot of unlearning, and this is difficult and painful, since it results in, I think, residual memetic debt, since an opportunity for reward is foregone because the transaction is now changed and the reward is no longer forthcoming.
What about attempts to integrate a foreign memetic system and the resulting residual memetic debt? From the position of a victim of the oppression, one makes overtures, one deploys memes with the expectation of reward, and that reward is not forthcoming, especially in a racist, sexist, whateverist regime, or an oppressive meme is deployed, an immunomeme. That is the mode, the basic concept of transforming a memetic system to integrate another cohort's memetic system. What is the objective, after all? To allow people to be different and get treated the same? To allow people to be different, to do different things and get the same response as others, or evince novel responses, i.e. new memes?
How can these this work?
The deployment decision algorithm tends to prefer LOW transaction granularity. Integrating a new subcohort's memetic system, you need to add memes to the main memeplex for special interactions with the incursive subcohort. What do these look like?
There's a "mother-in-law language" type of phenomenon in classist interactions, racism included. There are modes of address that bridge class boundaries, and different vocab. These may manifest themselves as "politeness levels" or other such. But is this a fundamental, sino quod non factor in memetically delineating class boundaries? These may be clothing, splitting on, built-in violence associated with incursions, physical, verbal or other, into another class' territory.
These things must be dismantled, which mean that means that memetic rewards are taken away. In principle these must be replaced. Otherwise, there is residual memetic debt, and this can quickly lead to violence and other upsets. Again, transaction granularity rears its head.
There’s a huge amount of transactional compaction around race and class interactions, built up over time, because these people don’t want to interact, they don’t really want to think about one another in any “complex” way. Memetically, the problem is that they only have a limited number of contact memes to interface across class/race boundaries, and with one another about other classes/races.
There are set memetic rewards associated with these interactions, usually associated with the oppression of other classes/races. These rewards must be somehow replaced, and without decreasing the transaction granularity overmuch. That is the issue for the majority wanting to move towards acceptance of some majority with a different culture. The majority may or may not be okay with having four new kinds of bathrooms, for instance. But in order for them to accept it, they need to have a new set of memes which will garner them rewards when using them in interactions with other majority members, and if at all possible, there should not be an enormous number of new memes, words, MIAOs that they have to learn to do it.
 Of a common cohort, sharing a relatively larger number of memes with one another, of course.