faces proceeds, while generating different face categories. At six months, a pattern of perceptual narrowing emerges, whereby a category of familiar faces is favored (Pascalis et al. 2002; cf. also Bower 1966 and Table 2.2). A comparable ontogenetic phenomenon of narrowing the perceptual window is also known from auditory perception. Young infants distinguish the phonemes of different languages much better than do infants from 8 months of age, who focus on the native language.
Table 2.2 From heterogeneous summation to Gestalt-perception in 24 babies at different ages (n = 6/age). (Modified after Bower 1966).
In human Gestalt perception invariance plays a role. In e-mails, the configurations (:-) and (:-/ are correctly interpreted despite their sideways orientation. We recognize the letter B independently of boldface, contrast, italics, fonts, or size (Table 2.1). However, invariances are not unlimited: depending on orientation and context, it can be interpreted differently.
Table 2.1 Invariance in Gestalt perception and its limitation.
An example of invariance in auditory perception concerns melodies. We recognize Mozart’s “Eine kleine Nachtmusik” independently of the key in which it is played or its instrumentation or whether it is whistled or sounded on a comb.
The Relational/Combinatorial Principle of Sign-stimuli from Other Sensory Modalities
The notion that the releasing values of sign-stimuli often depend on characteristic relationships between features accords with examples from other sensory modalities, such as chemical or auditory signals used in communication.
Female moths use sex-attractant pheromones to invite their males. In many moth species these consist of blends of two chemical compounds (Kaissling 2014). Whereas the same two components can be emitted by phylogenetically related species, a species-specific signal is produced by a characteristic mixture: “principle of parsimony.” Among species of the North American female leaf-roller moth the proportion of the pheromone compounds [Z]-11-tetradecenyl-acetate: [E]-11-tetradecenyl-acetate is:
90:10 in Archips mortuanus
60:40 in Archips argyrospilus
17:83 in Archips cerasivoranus
In various species of frogs, to choose another example, the females are attracted by the advertisement calls of the conspecific males if certain low-frequency and high-frequency components of sufficient energy in the call power-spectrum coincide (Capranica 1976):
200 Hz and 1400 Hz, bullfrog, Rana catesbeiana
500 Hz and 1500 Hz, leopard frog, Rana pipiens
900 Hz and 3000 Hz, green tree frog, Hyla cinerea.
Male adult bullfrogs among themselves evoke advertisement calls in a chorus. However, if young males start to take part, the adults become mute. Although immature, the young display sort of advertisement call, albeit a bit unsounded: high-frequency peak at 1400 Hz, but the low-frequency peak not at 200 Hz rather around 600 Hz. This 600-Hz component inhibits adult male’s mating-calling and female`s interest to follow. With age the vocal cavities of young males increase, so that the low-frequency peak shifts perfectly to 200 Hz (Capranica 1976). Hence, females are protected twofold from meeting an immature partner.
Unlike Rana or Hyla, the advertisement call of the Puerto Rican male treefrog, Eleuterodactylus coqui, is made up of a sequence of two notes with different conspecific addressees: the “Co”-note of 1100 Hz is addressed to males in male–male territorial interactions; it is followed by the “Qui”-note, upward sweeping from 1800 to 2100 Hz, to attract females (Narins 1981). The advertisement call of male tungara frogs, Engystomops pustulosus, too, consists of a sequence of different notes: a whine followed by several lower-pitched chucks (Ryan & Rand 1995). Whereas the whine serves species recognition, the chucks enhance the call’s attractiveness to females. Males of the related species E. coloradorum whine but do not chuck. If given a choice, their females prefer E. pustulosus calls (Chapter 12).
Enhancing effects may also result from combining features of different sensory channels. When prey is difficult to carry for Aphenogaster ants, they recruit help from workers by emitting a pheromone. If food competitors are present and time is short, ants also deliver a vibrational stimulus. Vibration alone has no effect on worker recruitment, but in combination it enhances the pheromone’s efficacy (Markl & Hölldobler 1978). Workers of leaf-cutting ants, Atta cephalotes, respond to a pheromone or vibration with recruitment behavior. If these ants are given a choice, they choose the combination, suggesting a kind of summation effect.
From Stimulus Summation to Supernormal Stimuli
Heterogeneous summation
The concept of Gestalt implies that the efficacy of a configuration is greater than the sum of the efficacies of its components (features). However, there are examples showing that independent different stimulus features are additive in their efficacy, whereby the whole is equal to the sum of its parts (Seitz 1940: “Reizsummenphänomen”). For instance, herring gulls, Larus argentatus, recognize their eggs by different features, such as size, shape, color. These features are additive in their influence upon retrieval of an egg having rolled out of the nest. Heiligenberg and coworkers (see Leong 1969) quantitatively demonstrated a comparable phenomenon in male perch, Haplochromis burtoni. The level of aggression A in these fish averaged at x bites/min. In a model fish (a), a black eye-bar increased the attack rate of conspecific males at
Aa = x + 2.8 bites/min,
while a model fish (b) with orange spots on skin, but no eye-bar, lowered the attack rate:
Ab = x–1.7 bites/min.
A model fish (c) containing eye-bar and orange spots caused an attack rate of
Ac = x + 1.1 bites/min,
which correlates well with the algebraic sum of the rates (1.0) obtained in (a) and (b). Hence, the opposite effects of both features summed algebraically: “heterogeneous summation.”
It is suggested that Gestalt perception of human faces is derived from heterogeneous summation both in phylogenetic and ontogenetic histories. For example, Bower (1966) showed in 2-month-old human babies that the number of conditioned orienting responses toward a face-model—consisting of head-outline, eye-dots, and cross—equaled the algebraic sum of the orienting activities measured to each component of the model (Table 2.2). In 5-month-old babies the responses to the face-model were about twice as high as the sum of the responses to each face component: Gestalt perception.
Supernormal stimulus
Stimulus summation introduces the phenomenon of supernormal stimulus, i.e., a stimulus that is more effective at eliciting a response than the stimulus for which it evolved. In studies using dummies, various examples show that exaggeration of a sign-stimulus leads to an extraordinary increase in its efficacy. The courting behavior of a male stickleback depends on the swollen abdomen of the pregnant female. When presented with two model females, one showing a normal swollen and the other a hyper-swollen abdomen, respectively, the male will choose the latter (Rowland 1989).
Another example concerns the egg-retrieval behavior of brooding greylag geese, Anser. After an egg rolls out of its nest, the goose extends its neck toward the egg and with its flat mandible slowly