2.1. Patients with a disproportionate deficit of plant knowledge

The findings with these patients are summarized in Table 1.

Table 1. Cases affected by selective or disproportionate impairment of plant life knowledge

	EA (Laiacona, Barbarotto, & Capitani, 2005)	Male	Herpetic encephalitis/left T (anterior and middle), and minimal right T
	ELM (Arguin, Bub, & Dudek, 1996)	Male	Stroke/bilateral medial T
	FAV (Crutch & Warrington, 2003)	Male	Stroke of left PCA/left T and O
	GR (Laiacona, Barbarotto, & Capitani, 1993)	Male	Head injury/left F–T
	JV (Pietrini, Nertempi, Vaglia, Revello, Pinna, & Ferro-Milone, 1988)	Male	Herpetic encephalitis/left: anterior and medial T, basal F, insula; right: insula
	KQu (Goldenberg, 1992)	Male	Stroke of left PCA/left T and O
	MD (Hart, Berndt, & Caramazza, 1985)	Male	Stroke/left F and basal ganglia
	MO (Cornil & Pillon, 2003)	Male	DAT/cortical and subcortical atrophy (no specific area implicated)
	RS (Samson & Pillon, 2003)	Male	Stroke of left PCA/left T and O
	SRB (Forde, Francis, Riddoch, Rumiati, & Humphreys, 1997)	Male	Artero-venous malformation/left: inferior-medial T, O; right: thalamus
	TU (Farah & Wallace, 1992)	Male	Artero-venous malformation/left O

We report sex, etiology, and lesion site. F, frontal; T, temporal; O, occipital; PCA, posterior cerebral artery; DAT, dementia of Alzheimer type.

Case EA (Laiacona, Barbarotto, & Capitani, 2005) was initially affected by a global semantic deficit of biological categories as a consequence of herpetic encephalitis. After 10 years, his deficit became minimal for animals, while it remained severe for fruits and vegetables without any further change. The discrepancy between animals and plants was statistically significant at word–picture matching and verbal probes where nuisance variables and difficulty were controlled for.

Case ELM (Arguin, Bub, & Dudek, 1996) was referred to as a category-specific visual agnosic patient. The authors conclude that he was more defective with fruits and vegetables than with animals on picture naming.

Case FAV (Crutch & Warrington, 2003), besides a global anomia, developed a severe difficulty with fruits and vegetables on naming and comprehension tasks and on tests probing semantic knowledge. Word frequency and familiarity were controlled for, and some comparative subjects were selected to control task difficulty.

For case GR (Laiacona, Barbarotto, & Capitani, 1993), the data suggest a greater sparing of animals compared with fruits and vegetables in word–picture matching and on semantic probes.

Case JV (Pietrini, Nertempi, Vaglia, Revello, Pinna, & Ferro-Milone, 1988) presented an overall deficit for natural categories naming and verbal comprehension, and the authors claimed that he was more severely impaired for plants than for animals.

Case KQu (Goldenberg, 1992), on neuropsychological investigations, named fruits and vegetables consistently worse than any other category examined (including animals) both from line drawing and from description.

Case MD (Hart et al., 1985) was impaired with the naming of fruit and vegetable pictures; moreover, he was hesitant when asked to give judgments about the properties of a fruit.

Case MO (Cornil & Pillon, 2003) suffered from a probable dementia of the Alzheimer type. At the beginning of his cognitive deterioration, he was impaired on naming and on word–picture verification with plants. Later (second and third assessment within 18 months), his impairment extended to manufactured objects, but animals were always spared. Stimuli were closely matched for word frequency and sex-specific concept familiarity, but no further data nor any structured statistical analysis was given in the published report. (Thanks to a courteous communication by the authors, we may report further unpublished data on this case.) The plant deficit of MO was certainly of a semantic nature and was also evident in other tasks based on tightly controlled material, e.g., a property-verification task and a naming-to-definition task. Interestingly, MO's deficit was homogeneous for different plant subcategories (fruits, vegetables, flowers, etc.), whereas the patient was neither affected by deficits of color perception nor by a deficit of color knowledge and on categorization tasks he was confused only with fruit and vegetable pictures.

Case RS (Samson & Pillon, 2003) showed a disproportionate naming and conceptual impairment for fruits and vegetables; animals were at intermediate level. Statistical comparisons were carried out taking into account nuisance variables. Neither sex specific familiarity nor a loss of object-color knowledge was at the origin of his fruit and vegetable impairment. He was flawless on object-decision task.

Forde, Francis, Riddoch, Rumiati, and Humphreys, 1997 described case SRB who showed a deficit for natural categories and was more severely impaired on naming of fruits and vegetables than of animals (both on picture naming and on naming to definition), consistent across different tasks.

For case TU (Farah & Wallace, 1992), the authors envisaged a selective naming deficit for fruits and vegetables.

2.2. Patients with a disproportionate deficit of animals knowledge

Deficits limited to the category of animals are exceptional in the literature (Table 2): only three such cases have been described.

Table 2. Cases affected by selective or disproportionate impairment of animals knowledge

	EW (Caramazza & Shelton, 1998)	Female	Stroke/left F and basal ganglia
	LA (Gainotti & Silveri, 1996, Silveri & Gainotti, 1988)	Female	Herpetic encephalitis/bilateral inferior T, more severe on the left
	LH (Farah et al., 1989, Farah et al., 1996)	Male	Head injury/right T and F, left subcortical O–T, bilateral P–O

We report sex, etiology, and lesion site. F, frontal; T, temporal; O, occipital; P, parietal.

Case EW (Caramazza & Shelton, 1998) presented very strong evidence of a selective animal deficit. Two other cases, LH (Farah et al., 1989, Farah et al., 1996) and LA (Gainotti & Silveri, 1996, Silveri & Gainotti, 1988), will be included, although the reports were not totally exhaustive.

2.3. General comments on patients' data

The frequencies are small, but we find 11 of 11 males in Table 1 and 1 of 3 males in Table 2: the comparison via Fisher's exact test was significant (p=.033). This evidence suggests that males are more likely than females to present a deficit for plant categories, but not for animals. As such, it seems conservative to suggest that the prevalence of males among patients with a deficit of biological categories (not better distinguished) may depend on the frequent presence of plant stimuli in the test batteries. Regardless, the apparent male vulnerability with the category of plants (or a female advantage for the same category) calls for interpretation.

2.4. Sex-by-category interaction in normal subjects with reference to animal and plant categories

A variety of tasks have been used to study differential semantic knowledge among normal subjects. Picture naming studies examine a fixed and predetermined restricted vocabulary. Semantic fluency tasks examine an unrestricted vocabulary for a specific category in a restricted time span (typically 1 min) and are sensitive to differences among categories in actual vocabulary size. Object-reality decision is a visual perceptual task without verbal components: subjects are asked to judge whether pictures represent real or unreal things. The literature suggests that tasks using plant stimuli are performed better by females, although the same outcome was not observed with animals. A sex difference has been observed by several authors on different tasks (Table 3).

Table 3. Semantic category dissociations among biological categories reported for groups of normal subjects

	Task	Semantic categories investigated and sex asymmetry	Reference
	Picture naming	Female superiority for fruits and vegetables. Male superiority for animals	McKenna and Parry, (1994)
	Picture naming latencies	Females faster for natural categories (fruits, vegetables, and animals not analysed separately)	Laws (1999)
	Semantic fluency	Female superiority for fruits and vegetables. No difference for animals	Capitani, Laiacona, and Barbarotto (1999)
	Semantic fluency	Female advantage for fruits but not for animals	Laws (2003)
	Picture reality decision	Female superiority for fruits and vegetables. No difference for animals	Barbarotto, Laiacona, Macchi, and Capitani (2002)

Differential familiarity may be a confound. Albanese, Capitani, Barbarotto, and Laiacona (2000) asked a large sample of normal subjects to rate their familiarity with common objects belonging to six semantic categories. Different aspects of familiarity were assessed, but females consistently gave higher familiarity ratings to fruits, vegetables, and furniture items, whereas males showed more familiarity with tools. No sex discrepancies were observed for animals.

3.1. Does familiarity explain everything?

A closer look at the clinical data casts doubt on the possibility that familiarity plays a decisive role. Consider case EA (Laiacona et al., 2005, Laiacona et al., 1997). EA, 10 years after HSE, had recovered with animals but not with plants, and at the last examination a selective impairment of fruits and vegetables was evident. The authors used stimuli controlled for male-specific familiarity (Albanese et al., 2000); the latter values were used as a covariate in the analysis of word–picture matching and semantic probe questions, but a significant difference between plants and animals remained. Table 4 reports the mean scores of EA on a task of “pointing to picture on verbal command” and on semantic probes (a totally verbal task) regarding animals, fruits, and vegetables, and also reports the male familiarity with the items employed. EA's better performance with animal stimuli cannot be traced to any familiarity advantage as they were the least familiar items. EA did not have any special premorbid familiarity with animals. After his disease, he retired from work and helped his family, including shopping at the greengrocer, so that he was still in contact with plant objects. However, at the greengrocer, he used a written list prepared by relatives and chose from the shelf the objects whose names matched the words written on the list.

Table 4. Findings with case EA

	Category	Mean male familiarity	Word–picture matching	Semantic probes (n=60)
	Artifacts (n=30)	2.36 (S.D.=0.74)	93% (normal >83%)	98% (normal >94%)
	Animals (n=10)	1.44 (S.D.=0.24)	100% (normal >70%)	75% (normal >85%)
	Fruits+vegetables (n=20)	2.20 (S.D.=0.41)	35% (normal >90%)	52% (normal >91%)

Hits are reported for each task. Semantic probes were six for each stimulus. Familiarity of the stimuli was rated on a 1-to-5 scale (5=very familiar) by 105 males (Albanese et al., 2000). Artifacts included 30 stimuli (10 tools, 10 vehicles, and 10 pieces of furniture), and there were 10 fruit and 10 vegetable stimuli.

A warning against explanations based on familiarity was also given by Samson and Pillon (2003) in their discussion of case RS. Also in this case, the difference between plants and animals was confirmed using sex-specific familiarity values and using fruit and vegetable stimuli that presented the highest familiarity and the lowest visual complexity within the battery. Interestingly, RS's naming of fruits and vegetables contrasted with his naming of domestic implements, a nonliving set likely to match fruits and vegetables for idiosyncratic familiarity since SR stated that he was poorly acquainted with household activities. The patient was impaired with fruits and vegetables but not with domestic implements.

On the basis of the analysis of these two cases, the naïve explanation that males are more impaired with fruits and vegetables because they are not acquainted with shopping at the greengrocers or with cooking seems rather unconvincing. We can also add that a greater premorbid familiarity with a given category seems insufficient to prevent a category from being severely affected by semantic disorders. For instance, case Michelangelo (Sartori, Job, & Coltheart, 1993), who was affected by a severe disorder of the semantic knowledge of animals (not in Table 3 because he was not tested for semantic knowledge of plants), had been an active member of the World Wildlife Fund, and in his premorbid life was acquainted with a vast number of animals. Another case anecdotally cited by Humphreys and Forde (2001) is that of a food expert and wine connoisseur who nevertheless presented a deficit in naming fruits and vegetables after brain damage. Recently, Humphreys and Riddoch (2003) reported a patient (GA) suffering from herpetic encephalitis who, besides an impairment with living categories, also presented a semantic deficit for musical instruments, notwithstanding the fact that he was a former professional musician. As these reports cast doubt on the familiarity account, we should consider alternative explanations.

3.2. Can neuropathology explain why males are more defective with fruits and vegetables?

If we neglect the category-by-sex interaction observed in studies of normal subjects, an explanation could be sought on purely neuropathological grounds. We could reproduce the observed sex by category interaction if the disproportionate deficit of fruits and vegetables were strongly linked to a given lesion site, and if diseases damaging this site were more frequent among male patients. This account is suggested by the evidence that three cases (FAV, KQu, and RS) of plant life impairment (Table 1) suffered from posterior cerebral artery (PCA) stroke that caused unilateral left posterior temporal and occipital damage. However, only three cases out of 11 were affected by PCA strokes. Among the remaining cases, as reported in Table 1, the lesion site did not invariably coincide with the areas supplied by the PCA.

3.3. Are sex differences in processing plant life items supported by neuroanatomical data?

Neuropathological explanations based on a different lesion site between males and females cannot explain why normal subjects show the same sex effect, with a female advantage for fruit and vegetable stimuli. Moreover, the anatomical basis of category specificity is still debated (Martin & Caramazza, 2003). Some biological objects (faces, animals, and plants) typically show peak activity in the lateral portion of the fusiform gyrus; however, the available functional imaging studies do not indicate whether the cognitive representation of animals and plant life is the same or different, and have not investigated whether the hypothetical substrate of these representations coincides in males and females. Therefore, presently, it seems wise to discuss this topic at the level of the behavioral pattern observed in clinical studies rather than at the underlying structural level.

3.4. A reappraisal of familiarity in evolutionary terms

A naïve familiarity account would suggest that acquired cognitive representations of plant exemplars are more robust among females because in their lifetime they have accumulated greater experience with these items. In turn, this implies that the different social roles of males and females may so strongly influence our cognitive system as to create a clinical discrepancy in the frequency of particular neuropsychological disorders, as well as produce parallel effects detectable among normal subjects. But this differential familiarity may have considerable evolutionary time depth (Silverman & Eals, 1992, Zihlman, 1981) and thus may have exerted evolutionary as well as ontogenetic effects on cognitive processes. Females may have specialized on plant resources (and males on animal resources) for many hundreds of thousands of years. But while the evolutionary account can explain the female advantage for plant knowledge, it does not necessarily imply a complementary female deficit for animal knowledge. Humans' interactions with animals may have been as predators or as prey. The latter are clearly relevant to fitness and probably impacted both sexes, and may have overshadowed any pressures arising out of male hunting activity.

The hypothesis that the female advantage for cognitive representation of plants has an evolutionary origin was introduced by Barbarotto, Laiacona, Macchi, and Capitani (2002) and was reappraised by Laws (2003). This account is consistent with a wider theoretical framework able to interpret data from different fields, e.g., female advantage on memory tasks (Gaulin et al., 2005, McBurney et al., 1997, Silverman & Eals, 1992) where familiarity is unlikely to play a role. At a broader methodological level, the possible associations between pathological conditions and normal patterns of performance offer a useful approach to the study of evolved cognitive competencies and the mechanisms that support them.