It is interesting to me that some theorists (who work with DID) consider BPD to be a dissorder that is best thought of on a continuum with DID (so that BPD is thought of as a dissociative / trauma disorder) e.g., Colin Ross. While other theorists (who work with BPD) consider DID to be a disorder that is best thought of on a continuum with BPD (so that DID is thought of as a disorder of emotion dysregulation).

One could say that they are both right in the sense that they are indeed on a continuum.

People on the BPD end of things are likely to consider alters the product of therapy, however. People on the DID end of things are likely to consider alters to have been present before therapy.

Dennett (interestingly enough) looks at what a self is supposed to be... Seems that the self is a social construction. The social construction is shaped by models in our life, the narrative stories we tell ourselves, reinforcement from others. He considers alters to be narrative stories comperably to how a single self is a narrative story.

Whether the notion of multiple selves / personalities / alters occurs to the client at the suggestion of an abuser, at the clients own suggestion, at the suggestion of movies or books, at the suggestion of a therapist doesn't really matter. He thinks it is hopelessly arbitrary to consider some of those aetiologies to be somehow 'legitimate / genuine' while others are not.

What it does draw ones attention to, however, is the notion that behaviours are indeed (unconsciously) under the control of reinforcement contingencies and therapists are the source of reinforcement contingencies.

IMO... There is an unjustifyable difference between therapists considering some clients to be 'attention seeking' and 'manipulative' in a disorder that is likely to last a lifetime while therapists consider other clients to have a disorder that is 'more severe' yet the treatment outcomes are thought to be 'much better'.

IMO 'attention seeking' and 'manipulative' are expressions of a therapists lack of understanding / feelings of inadequacy.

IMO 'alters' are narrative constructs that are employed to 'make the best sense of behaviour' where behaviour comes under the control of reinforcement contingencies that therapists do indeed have a significant role in shaping.

I advocate a middle ground... Selves just are narrative constructs so there is no metaphysical puzzle in the notion of many selves that isn't already present in the notion of a single self. Punishment isn't particularly effective in shaping behaviour, but teaching legitimate alternatives that have the same function (much needed attention and care) is much much better in shaping behaviour so that a single systems (self) interpretation is more appropriate.
Disclaimed actions should not absolve someone of consequences for their actions. Inappropriate actions need to be dealt to to make their reoccurance less likely.
But the latter point probably follows from my notion that just because we 'could not have done otherwise' (in the sense that determinism could well be true and even if indeterminism was true then the probabilities of different actions could not have been otherwise) and hence justice / moral responsibility is about delivering consequences to make future problem behaviours less likely.