Abstract: Some claims about poems are uncontroversial: that a poem is composed in dactylic hexameter, as in Homer’s epics, or in iambic pentameter, as in Shakespeare’s sonnets, or in no particular traditional meter, as in most of e. e. cummings’s work; that it rhymes following an abab pattern or that it does not; that it is very long, very short, or any length in between; that it employs sophisticated diction, archaic language, or common everyday words; that its similes and metaphors are novelor clich ?e. Such claims may easily be ascertained by those able to count syllables, those able to distinguish the stressed syllables from the unstressed dones, and those familiar with the varieties of poetic meter; by those able to tell whether two or more words sound alike; by those able to distinguish different text lengths; by those able to recognize when words are of the garden-variety kind. Except for familiarity with the kinds of poetic meter, ‘those’ are most of us. We may call these ‘base,’ ‘lower level,’ or ‘structural’ properties. At another level of description, poems may be tightly knit, unified, balanced, heavy and somber, light and jolly, and so on. The attributions in this case are still descriptive, but an evaluative judgment may be embedded in them, or it may be typically taken to be embedded, or intended to be embedded, in them. That is, a positive or negative valence may sometimes accompany the judgment that a poem is unified and balanced: one may find it good or bad in virtue of those characteristics (though one may also find it good or bad regardless of those characteristics). Further, justifications of why a poem is unified, balanced, and so on are made by reference to the qualities specified at the first level of description: it is unified because all the parts fit well together in some manner. We may call these ‘aesthetic attributions’; as Jerrold Levinson puts it, here we have ‘an overall impression afforded, an impression that cannot simply be identified with the structural properties that underpin it.’ At a third level still, and in part in virtue of facts at the previous two levels, poems may be beautiful, terrific, or horrendously bad: here we have wholly evaluative attributions, or aesthetic judgments properly so-called. Note how there isan inverse proportion in informative value be-tween base properties and aesthetic judgments: base properties are informative about a work (‘is in iambic pentameter’) but not aesthetically evaluative and thus not aesthetically informative; aesthetic evaluations (‘is beautiful’) are aesthetically informative, but tell us nothing about the specific characteristics of a work. Aesthetic attributions fall in the middle also in that they may retain some of the informative value of either extreme: they may be somewhat structurally informative and some what aesthetically informative (‘unified’). If it is true that these three levels are at once distinguishable and intrinsically related, some questions one may ask are: How are they related? How is our perception of a set of words arranged in a certain cadence and with breaks visually or aurally marked related to our perceiving in the mor attributing to them a certain set of aestheticqualities? How do we go from characteristics such as ‘has lines of eighteen syllables, where a marked syllable is followed by two unmarked ones throughout’ to ‘is tightly knit’ to ‘is beautiful’? In other words, how do we move from purely descriptive attributes to aesthetic and evaluative ones? Anyone may count syllables, and most of us can more or less tell when a syllable is stressed relative to another that precedes or follows it. We may likewise be able to judge whether a metaphor is unusual or not simply by recalling whether we have heard anything like it in the past, or how unlike each other the terms of comparison are. That assessment may be accompanied by approval or disapproval; in itself it need not express either (‘That’s a novel metaphor: it is awful’ is a perfectly sensible statement). Finally, when we move to ‘beautiful’ and ‘moving,’ we are making a judgment of taste: our approval is embedded in those terms. My concern in what follows is with the move from lower-level perceptual qualities to the attribution of aesthetic qualities. I am not concerned with how we go from there to an overall aesthetic evaluation. In my proposal, I question the much discussed wisdom handed down to us by Frank Sibley. I am referring to Sibley’s famous claim, defended in ‘Aesthetic Concepts’ and related articles, that we are never, in any art form, warranted in making the (logical) jump from the description of non aesthetic properties to the ascription of aesthetic ones. In his words, Sibley claimed that ‘there are no non aesthetic features which serve in any circumstances as logically sufficient conditions for applying aesthetic terms.’ We cannot, for instance, go from ’employ[s] bright colors’ to ‘is lively and vigorous,’ the way we can go from ‘unmarried male’ to ‘bachelor’ or from ‘enclosed figure with four equal sides and four right angles’ to ‘square.’ Surely we cannot, but why should anyone have thought otherwise? Aesthetic qualities are qualities, not concepts. As an attribute, ‘graceful’ more closely resembles ‘hot’ than it does ‘square.’ There is no reason to expect a one-to-one relationship between base properties and aesthetic attributions, but there is good reason to expect that a range of properties is clearly associated with a range of attributions, just as a range of temperatures is associated with feeling cold. Sibley also claimed that no particular base property or set thereof is necessary for any given aesthetic concept to apply. This is because things may have the same aesthetic quality for different reasons: ‘one thing is graceful because of these features, another because of those, and so on almost endlessly.’ I do not question whether Sibley’s claims are defensible when it comes to vases, paintings, sculptures, or sonatas; indeed, his view is compelling as a general rule. However, it seems to me that some varieties of poetry provide, not an exception to Sibley’s rule – I am not claiming logical entailments here, nor do I think any- one could – but evidence for what may be called a ‘defeasible guarantee.’ In at least some kinds of formal poetry, there is a sense in which a description in nonaesthetic terms sometimes ought to suffice, in virtue of what we may call ‘psychoaesthetic’ associations between the perception of formal features and felt aesthetic qualities, for the attribution of an aesthetic quality. Accordingly, my first goal in what follows is to show in what way I think it is sufficient and to provide some examples in support of that connection. I hope that from this it emerges that Sibley was wrong to hold that unless their relationship is a logicoconceptual one, no base properties ever suffice to warrant the ascription of an aesthetic quality.