What an excellent question, projective identification is a very complex defense. I have often felt that projective identification is one of the hidden geniuses of Melanie Klein and the British object relations movement because it is a defensive process that suggests a two or more person psychology (which has been growing in popularity as a modern paradigm shift in psychoanalysis, sometimes referred to as the relational perspective).
When looking at the card images alone, I think of any of the cards that show more than one individual (e.g., in the RWS stylized cards the 5 of swords, 10 of pentacles, The Lovers, 5 of wands, 6 of wands, 6 of pentacles etc.) which can also take a more positive slant (e.g., 2 of cups, 3 of cups, 4 of wands, 10 of cups, etc.). However, I believe context is everything when it comes to projective identification and that projective identification would have to be hinted at by the querent or client's perceptions of the interrelationships between individuals depicted on these cards more than anything else.
When allowing querents to project onto card images, especially when they reference themselves in relation to their perceptions of the images, I believe it is then possible to observe projective identification in virtually any card image. In such a case I beleive they are engaging in a relationship with such a card, thus they can be engaging in projective identification with the image. Again, I believe it is the context in which a person responds to a card image that will suggest whether projective identification is being used.
Above all I beleive a context of one's relationship has to be expressed while viewing the card(s), whereas referencing an image without this context brings little chance of uncovering projective identification.
(Although if I had to lay a bet on which card will most likely illicit this defence, my vote goes for the 5 of Swords.)
Forrest
