In this paper, I argue that construction is not merely a tool amongst the many that are at the clinician's disposal; construction is the very means and end of the psychoanalytic experience. I reinterpret Freud's classic notion in light of the Winnicottian found object and the Deleuzo-Guattarian understanding of desire as production. I then go on to deploy two detailed clinical vignettes in order to show how, thus reinterpreted, construction is at its most valuable when it is generative rather than historical. Instead of merely eliciting an accurate representation of a lost experience, construction brings about further associations, interpretations and, indeed, constructions that reshape the subject.