Spider heuristics

Fiona R. Cross^a, ,  and Robert R. Jackson^a, b, , 

^aSchool of Biological Sciences, University of Canterbury, Private Bag 4800, Christchurch, New Zealand

^bInternational Centre of Insect Physiology and Ecology (ICIPE), Thomas Odhiambo Campus, P.O. Box 30, Mbita Point, Kenya

Available online 11 March 2005. 

Abstract

Simple heuristics may help explain how even a spider, despite its minute brain, can be disturbingly intelligent. Hutchinson and Gigerenzer suggest that the generalist–specialist distinction (or more accurately the predictability–unpredictability distinction) may be related to a species’ level of reliance on simple heuristics, and spider behaviour may present some especially instructive opportunities for investigating these ideas. Daniel Dennett's distinction between Darwinian, Skinnerian and Popperian animals might be useful for discerning the different contexts in which optimality considerations and individual decision making are relevant.

Keywords: Spider; Salticidae; Portia; Predictability-unpredictability distinction

Article Outline

References

Hutchinson and Gigerenzer (2005) have given us a lot to think about, and we welcome the common ground they are trying to establish between what biologists (especially ecologists) call “rules of thumb” and what psychologists call “simple heuristics”. There is much remaining to be learned about how decisions are made by real people and real animals. ‘Rule of thumb’, like ‘simple heuristic’, seems to be a term chosen deliberately to discourage the notion that individuals do anything especially elaborate when making decisions, but biologists and psychologists put different spins on their respective terms. In biology, the expression ‘rules of thumb’ seems to be a spin-off from optimal foraging theory (OFT), almost as though OFT is apologising for the animal's limited brain power. The ABC group has a different perspective on simple heuristics, envisaging them as nothing to apologise for because, remarkably, simple heuristics can actually outperform more elaborate problem-solving procedures.

Our own research, as biologists, is based on using jumping spiders (Salticidae) to study decision making, problem solving, selective attention and other topics related to animal cognition (Cross and Jackson, in press; Harland and Jackson, 2004). Perhaps there is some common ground between understanding how simple heuristics can make us smart and understanding how salticids sometimes seem disturbingly intelligent for animals with minute brains. It is tempting to think that a big-brain animal like Homo sapiens could almost choose whether to use simple or not-so-simple heuristics, whereas the small-brain spider would have no alternative to adopting simple heuristics. Fortunately, simple heuristics can make even a spider smart. However, Hutchinson and Gigerenzer avoided the tired old clichés about small brains that condemn arthropods to rigid instinct-driven behaviour.

The distinction that Hutchinson and Gigerenzer make between specialists and generalists is interesting. What we encounter in our lives as humans appears to be at the top end for any species in terms of unpredictability. For people, it seems to be not only the ability to learn through experience and generalise across situations that is important. We routinely confront novel problems unlike what our ancestors faced during our evolutionary history. Being fast, frugal and effective, simple heuristics may be critically important for solving the vast spectrum of problems that challenge us. Investigating the generalist–specialist distinction (or more accurately the predictability–unpredictability distinction) and a species’ level of reliance on simple heuristics may be a particularly productive direction for future research by behavioural biologists.

It may be relevant for understanding spider behaviour. For most web-building spiders, life may be far more predictable than for Portia, a genus of tropical salticids that single out other spiders as preferred prey (Jackson and Pollard, 1996). We might call Portia a specialist at preying on other spiders, but another spider's web is for Portia an especially dangerous and unpredictable arena for predator-prey encounters. A spider's web is part of its sensory system (Witt, 1975), and Portia appears to gain dynamic fine control of the other spider's behaviour by adjusting the signals made in the web (aggressive mimicry) (Tarsitano et al., 2000). For example, many spiders may be lured in until close enough for Portia to attack, but Portia may avoid luring in long-legged pholcid spiders, choosing instead to keep the pholcid agitated and turning about in the web until there is a gap between the legs for a clear shot at the pholcid's body (Jackson and Wilcox, 1998).

Portia preys on many different kinds of spiders, and there appears to be considerable unpredictability in the particular tactics and signals that work best during encounters with spiders. Often Portia may use something like a take-the-best heuristic by generating different signals on the web until one of these signals elicits an appropriate response from the web resident and then repeating the successful signal for as long as it works (Jackson and Carter, 2001). Portia may even apply this same simple heuristic to a novel confinement problem, how to escape from an island surrounded by water (Jackson et al., 2001). However, we have little understanding of the heuristics by whichPortia sets its criterion for “best” in different encounters (e.g., how does Portia decide whether “best” is a signal that elicits approach or a signal that keeps the spider actively turning about, but not approaching?).

Where do optimality considerations fit into the scheme of things? Tinbergen's (1963) four explanation categories come to mind (see Dewsbury, 1999). “Rules of thumb” and “simple heuristics” appear to be terms for the “immediate causation” of an individual's decisions. Using OFT, elaborate fine tuning of behaviour (as an outcome of natural selection) can be suggested without necessarily arguing that individual animals do anything especially elaborate when they make their individual decisions. It may not be necessary to say they make decisions at all. For example, OFT has been used to explain ecotypic variation in the courtship persistence (giving-up time) of male salticids foraging for opportunities to mate with conspecific females. It was predicted that males from habitats where encounters with receptive females are more frequent would be innately less persistent in courtship than males from habitats where encounters with receptive females are less frequent. This prediction has been corroborated (Jackson, 1980), but the individual male salticid was never envisaged as somehow detecting the abundance of receptive females in the habitat and then using this information for making optimality decisions. The individual male's rule of thumb might be dead simple in this example (i.e., the individuals from any one population may simply adopt a set (innate) giving up time). The mechanisms underlying variation in the giving-up times of individual males might be something as simple as different thresholds for habituation (see Jackson, 1982). The male spider is what DanielDennett (1996) calls a Darwinian animal. It may be ‘intelligent’ in the sense of having a good solution to the problem of how long to persist in courtship, but natural selection, not the individual salticid, derived this solution. There is no need to envisage the individual male spider weighing up probabilities and doing optimality calculations.

A comparable generate-and-test algorithm (operant conditioning or trial-and-error learning) may enable individual animals to derive optimal solutions. Portia illustrates this when deriving signals by trial and error. This is what Dennett called a ‘Skinnerian animal’ and there has been a long history of avoiding terms like ‘reasoning’ when talking about Skinnerian animals.

Perhaps it is with Dennett's third kind of animal (Popperian) that we begin to admire the intelligence of the individual. By running something like a simulation in its head, a Popperian animal can forego the need for real trials in the physical world. Even spiders may be Popperian. Portia, for example, routinely plans and then executes detours by which it can sneak up on other spiders from behind (Jackson et al., 2002 and Tarsitano and Jackson, 1997). Although some kind of trial-and-error by simulation may underlie route choice, the particular route chosen need not first be tried out (Tarsitano and Andrew, 1999).

Optimality considerations give us insight into evolution and adaptation, but the question of how individuals make decisions becomes particularly interesting at the level of Popperian animals. Hutchinson and Gigerenzer have pointed the way toward testable hypotheses that might apply even to a spider.

References

Cross and Jackson, in press Cross, F.R., Jackson, R.R., in press. From eight-legged automatons to thinking spiders. In: Fujita, K., Itakura, S. (Eds.), Diversity of Cognition. Kyoto University Academic Press, Kyoto.

Dennett, 1996 D.C. Dennett, Kinds of Minds: Toward an Understanding of Consciousness, Basic Books, New York (1996) 184 pp..

Dewsbury, 1999 D.A. Dewsbury, The proximate and the ultimate: past, present and future, Behav. Processes 46 (1999), pp. 189–199. Article |  PDF (90 K) | View Record in Scopus | Cited By in Scopus (6)

Harland and Jackson, 2004 D.P. Harland and R.R. Jackson, Portia perceptions: the Umwelt of an araneophagic jumping spider. In: F.R. Prete, Editor, Complex Worlds from Simpler Nervous Systems, MIT Press, Cambridge, MA (2004), pp. 5–40.

Hutchinson and Gigerenzer, 2005 J.M.C. Hutchinson and G. Gigerenzer, Simple heuristics and rules of thumb; where psychologists and behavioural biologists might meet, Behav. Proc. 69 (2005), pp. 97–124. Article |  PDF (353 K) | View Record in Scopus | Cited By in Scopus (59)

Jackson, 1980 R.R. Jackson, The mating strategy of Phidippus johnsoni (Araneae, Salticidae): IV. Interpopulational variation in courtship persistence, Behav. Ecol. Sociobiol. 6 (1980), pp. 257–263. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (4)

Jackson, 1982 R.R. Jackson, Habituation as a mechanism of intersexual selection, J. Theor. Biol. 97 (1982), pp. 333–335. Abstract | View Record in Scopus | Cited By in Scopus (3)

Jackson and Carter, 2001 R.R. Jackson and C.M. Carter, Interpopulation variation in use of trial-and-error derivation of aggressive-mimicry signals by Portia labiata from the Philippines, J. Insect Behav. 14 (2001), pp. 799–827.Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (8)

Jackson and Pollard, 1996 R.R. Jackson and S.D. Pollard, Predatory behavior of jumping spiders, Annu. Rev. Entomol. 41 (1996), pp. 287–308. View Record in Scopus | Cited By in Scopus (64)

Jackson and Wilcox, 1998 R.R. Jackson and R.S. Wilcox, Spider-eating spiders, Am. Sci. 86 (1998), pp. 350–357. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (26)

Jackson et al., 2001 R.R. Jackson, C.M. Carter and M.S. Tarsitano, Trial-and-error solving of a confinement problem by a jumping spider, Portia fimbriata, Behaviour 138 (2001), pp. 1215–1234. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (6)

Jackson et al., 2002 R.R. Jackson, S.D. Pollard, D. Li and N. Fijn, Interpopulation variation in the risk-related decisions of Portia labiata, an araneophagic jumping spider (Araneae, Salticidae), during predatory sequences with spitting spiders, Anim. Cogn. 5 (2002), pp. 215–223. View Record in Scopus | Cited By in Scopus (13)

Tarsitano and Andrew, 1999 M.S. Tarsitano and R. Andrew, Scanning and route selection in the jumping spider Portia labiata, Anim. Behav. 58 (1999), pp. 255–265. Abstract |  PDF (229 K) | View Record in Scopus | Cited By in Scopus (18)

Tarsitano and Jackson, 1997 M.S. Tarsitano and R.R. Jackson, Araneophagic jumping spiders discriminate between detour routes that do and do not lead to prey., Anim. Behav. 53 (1997), pp. 257–266. Abstract |  PDF (340 K) |View Record in Scopus | Cited By in Scopus (22)

Tarsitano et al., 2000 M. Tarsitano, R.R. Jackson and W. Kirchner, Signals and signal choices made by araneophagic jumping spiders while hunting the orb-weaving spiders Zygiella x-notata and Zosis genicularis, Ethology 106(2000), pp. 595–615. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (15)

Tinbergen, 1963 N. Tinbergen, On the aims and methods of ethology, Z. Tierpsychol. 20 (1963), pp. 410–433.