Philosophers are never quite sure what they are talking about — about what the issues really are — and so… — Daniel Dennett

The trouble with the canons of scientific evidence [...] is that they virtually rule out the description of anything but oft-repeated, oft-observed, stereotypic behavior of a species, and this is just the sort of behavior that reveals no particular intelligence at all — all this behavior can be more or less plausibly explained as the effects of some humdrum combination of "instinct" or tropism and conditioned response. It is the novel bits of behavior, the acts that couldn't plausibly be accounted for in terms of prior conditioning or training or habit, that speak eloquently of intelligence; but if their very novelty and unrepeatability make them anecdotal and hence inadmissible evidence, how can one proceed to develop the cognitive case for the intelligence of one's target species?