The Street here follows the practice of several Weimar fllms that revolve around the motivated view (often not identical with the classic point-of-view shot) and the power of off-screen space, as a double structure that regulates, but also suspends the characters' positions within the film. In particular, off-screen space becomes the signifier of a source of power that forever escapes the protagonists' control. It creates the structure of excess already mentioned and which the point-of-view shots can never quite recover or contain. As in The Street and The Last Laugh, there always seem to be more pairs of eyes in a Weimar film than there are characters on the screen. While motivated views are very frequent, the reverse-field shot is in fact rather rare, and off-screen space retains its powers of suggestion, menace, dread. 68 One might generalize and say that what typifies this cinema is that power is equated with vision, and vision with knowledge, and knowledge with control, and control with anxiety in a power/anxiety/knowledge nexus almost entirely mapped on the axis seeing/unseen/being seen. The dynamic is thus circular, or rather, it functions according to interlocking gradients, rather like the children's game where scissors, stone and paper each has the power to defeat or neutralise the other. This very dialectic of motivated look, onscreen and off-screen space in the German cinema also determines the viewing positions of the spectator, and is responsible for the problematic status of the image, where one often has the impression that the image is both hyper-real (because invested with emotional significance) and unreal (because insufficiently articulated within a time-space continuum or a chain of cause and effect). Attention is directed towards off-screen space, emphasising what is not seen but must be there, and thus invoking a (symbolic) absent cause, invested by the spectator with imaginary power, encouraging him to give shape and substance to the invisible. Conversely, the sheer weight of this invisibility makes the space and action that the audience does see an always-already-seen-space, readable either as a visual configuration that 'belongs' to some other presence, or as an iconographic allegory that requires interpretation.69 As Janet Bergstrom has suggested, especially in relation to Murnau and Lang, narrative progress in Weimar cinema is achieved by discontinuity and gaps, binding but also 'trapping' the spectator in elliptical constructions, so that one of the filmic codes most relevant for narrative development is not seeing/seen, but seen/unseen, or seen/inferred, that is, the action advances across the imaginary space generated by the off-screen gaze. 70 Perception being so often the most fallible and deceptive of indices to truth, it can be questioned by and subordinated to off-screen space. Following Mary Ann Doane, one might say that in the German cinema, there is also a problem about object relationships: seeing/seen is differently valorised because the gaze so often does not have an object, or does not 'fetishise' its object. Instead of the reverse-field, point-of-view structure so central to classical narrative, one finds that no clear subject-object relations are allowed to develop, on which the narrative referents might be mapped. Rather, the field of vision

tends to lead to phobic or paranoid states, often connoted by the 'double', which is to say, giving rise to a narcissistic structure, where the subject appears in the place of an 'Other' along either a temporal or spatial axis of division and discontinuity. 71

Spectator positions are thus marked by the same lack of knowledge, the same anxiety that characterises the protagonists: torn between on-screen space, which seems insufficient and off-screen, which is ('dread-fully') unknowable, the spectator is locked into the fiction as a split subject. It is this uncomfortable, anxious subject position that makes Weimar films often seem to be a taxingly concentrated experience. American cinema of the 'teens, and especially the early films of Griffith were not without their own forms of non-continuity and tableau shots, but they avoided narrational ambiguity (Griffith, for instance, was not comfortable with off-screen space), even where the settings were stylised (as in the 'expressionistically' lit interiors of Ralph Ince and Cecil B. de Mille).72 Yet by the 1920s the German cinema had developed a narrative system that classical Hollywood had begun to marginalise,73 and this divergence between American and German practice may be why the German films have elicited so many divergent psychological, sociological and ideological-political readings, appearing incomprehensible unless made coherent thanks to an imagined referent (in the case of Kracauer, nascent Nazism; for others, the trauma of the First World War). Retrospectively identified, such historical referents must be secondary effects, in the eye of the beholder, because the films withhold such certainty. Kracauer's political thesis, for instance, requires that he interprets the films' special visuality and non-linear temporality as 'premonition'. He treats Weimar narratives as if the default value was the Hollywood norm, with a clear articulation of space and time, whose lack in Weimar films he interprets as a sort of future perfect of 'anticipation'. But the films put in crisis the very notion of a textual norm, along with a stable authority, a reliable narrator and a voice of truth. It would seem that the underlying indecidability of many of the key Weimar narratives are a further indication that what is at issue is a particular kind of engagement with the spectator, rather than a 'representation' of external reality, social or psychological, present, past or future. What one retains from the films is thus not the political tendencies one can extrapolate, but their resistances to such extrapolations, their constitutive hesitation or doubleness - in short, their investment in something other than, or apart from, narrative as the depiction of character-centred action.