Penn et al. review a number of papers that demonstrate claims for human uniqueness and then proceed to a theoretical stance supporting this view. Given word limitations, I merely present data showing why Penn et al. should take care when arguing for lack of specific nonhuman capacities and for strong discontinuities between human and nonhuman competencies.

I begin with their discussion of same–different. Most of Penn et al.'s cited references do indeed fail to demonstrate nonhumans' comprehension of the relation; however, Penn et al. miss data from the Grey parrot study (Alex, Psittacus erithacus) that make a stronger case for human-like competence. Alex, who could observe two completely novel objects and respond to questions of both “What's same?” and “What's different?” with the label of an attribute that was same or different between the items (“color,” “shape,” “matter”; Pepperberg Reference Pepperberg1987) or with “none” if nothing was same or different (Pepperberg Reference Pepperberg1988), demonstrated, in Penn et al.'s words, a categorical distinction between displays with no item variability and those with any item variability at all. Furthermore, Alex, without any training, when queried “What color bigger?,” transferred use of “none” to label absence of a size difference between two objects; thus he could transfer to an untrained format (Pepperberg & Brezinsky Reference Pepperberg and Brezinsky1991).

Penn et al.'s discussion of transitive inference in nature is also limited in scope. They seem unfamiliar with studies on female great tits (Parus major), who decide whether to enter a male neighbor's territory (probably for extra-pair copulations) after eavesdropping upon experimentally manipulated, unobserved interactions between a stranger (an experimental playback) and her mate and the same stranger and said neighboring male. To succeed, the female has to distinguish her various neighbors from her mate and retain what she knows about the relative worth of her mate M, and each neighbor N. She must identify a new male, S, briefly listen and determine his rank in a contest versus M, and then do the same in another contest versus N. She must store and compare these two rankings and then infer the relative ranking of M and N based on their rankings with S, possibly updating her stored original memory. That is, the female tit makes her decision by inferring the ranking of the two resident males based on their respective abilities in dealing with the same intruder, and is more likely to enter the territory of the neighbor if he is inferred to be dominant to her mate (Otter et al. Reference Otter, McGregor, Terry, Burford, Peake and Dabelsteen1999; Peake Reference Peake and McGregor2005). Interestingly, because the relative rankings of males chosen for the experiment were unknown and the choice was random as to whether a given playback would simulate a dominant or a subordinate interaction, the information might counter what the female tit knows about previous interactions between her mate and her neighbor. Thus she is not simply making “an egocentric prediction about how to respond to a potential rival,” but rather doing what Penn et al. argue is impossible for nonhumans.

A third case relates, albeit somewhat indirectly, to Penn et al.'s discussion of advantages that humans have on account of their capacity for symbolic communication as well as their ability to reason in a relational manner. Again, Penn et al. overlook research with the Grey parrot Alex (Pepperberg Reference Pepperberg2006). The study in question combined several of Alex's capacities. As noted earlier, he could label the color of the bigger or smaller object in a pair; he could also vocally quantify up to six-item sets (including heterogeneous subsets, i.e., quantify the number of blue blocks in a collection of blue and green balls and blocks; Pepperberg Reference Pepperberg1994). He was separately trained to identify the Arabic numerals 1 to 6 with the same vocal English labels, but was never trained to associate these Arabic numbers with their relevant physical quantities or to order the numerals or sets with respect to size. He was then shown pairs of Arabic numbers or an Arabic numeral and a set of objects and was asked for the color of the bigger or smaller numeral or set. Alex's high success rate showed he (a) understood number symbols as abstract representations of real-world collections, (b) inferred the relationship between Arabic numerals and their quantity via stimulus equivalence, and (c) understood the untrained ordinal relationship of his numbers. Obviously, Alex possessed skills not acknowledged by Penn et al. as being possible for nonhumans.

I wonder why Penn et al. fail to mention research by Timothy Gentner and colleagues (T. Gentner et al. Reference Gentner, Fenn, Margoliash and Nusbaum2006) demonstrating that recursive-style behavior, which Hauser et al. (Reference Hauser, Chomsky and Fitch2002a) claim is unique to humans, can be observed in starlings (Sturnus vulgaris). Now, many colleagues, including those cited by Penn et al., do not agree with the definition of recursion used by Hauser et al. nor do they believe that recursion is indeed the one defining factor that separates human and nonhuman communicative behavior. Nevertheless, Gentner et al. do provide evidence for an advanced nonhuman capacity that Penn et al. might well have discussed.

In sum, although Penn et al. do indeed present cases for which no good data as yet exist to demonstrate equivalent capacities for humans and nonhumans, I disagree with their insistence that the present lack of such data leads to a theoretical stance requiring a sharp divide between human and nonhuman capacities. Absence of evidence is not a sure argument for evidence of absence. A continuum appears to exist for many behavior patterns once thought to provide critical distinctions between humans and nonhumans; I discuss some such instances missed by Penn et al., others also exist, and I suspect that, over time, researchers will find more continua in other behavior patterns. Moreover, although I suspect that some of the papers that I cite were not published when this target article was written, their recent appearance only supports my point – that new data may require a reappraisal of purported certainties. One may argue about definitions of discontinuity – for example, how to reconcile some societies' advanced tool creation and use with those of primitive societies whose tools are not much better than those of corvids (Everett Reference Everett2005; Hunt & Grey Reference Hunt and Grey2007) – and I do not deny the many differences that indeed exist between humans and nonhumans, but I believe future research likely will show these to be of degree rather than of kind.