Review. Theoretical and empirical evidence for the impact of inductive biases on cultural evolution - PubMed (original) (raw)

Review

Review. Theoretical and empirical evidence for the impact of inductive biases on cultural evolution

Thomas L Griffiths et al. Philos Trans R Soc Lond B Biol Sci. 2008.

Abstract

The question of how much the outcomes of cultural evolution are shaped by the cognitive capacities of human learners has been explored in several disciplines, including psychology, anthropology and linguistics. We address this question through a detailed investigation of transmission chains, in which each person passes information to another along a chain. We review mathematical and empirical evidence that shows that under general conditions, and across experimental paradigms, the information passed along transmission chains will be affected by the inductive biases of the people involved-the constraints on learning and memory, which influence conclusions from limited data. The mathematical analysis considers the case where each person is a rational Bayesian agent. The empirical work consists of behavioural experiments in which human participants are shown to operate in the manner predicted by the Bayesian framework. Specifically, in situations in which each person's response is used to determine the data seen by the next person, people converge on concepts consistent with their inductive biases irrespective of the information seen by the first member of the chain. We then relate the Bayesian analysis of transmission chains to models of biological evolution, clarifying how chains of individuals correspond to population-level models and how selective forces can be incorporated into our models. Taken together, these results indicate how laboratory studies of transmission chains can provide information about the dynamics of cultural evolution and illustrate that inductive biases can have a significant impact on these dynamics.

PubMed Disclaimer

Figures

Figure 1

Figure 1

Transmission chains provide a simple setting for studying cultural transmission that has been used in psychology, anthropology and linguistics. In a transmission chain, each agent observes the data generated by the previous agent, forms a hypothesis about the source of these data and then uses that hypothesis to generate data for the next agent.

Figure 2

Figure 2

Dynamics of the probability of an agent adopting hypothesis 1 as a function of the number of generations of transmission. As the number of generations increases, the probability of choosing _h_1 converges to the prior probability, _π_=0.2. The noise parameter ϵ determines the rate of convergence, with _ϵ_=0.01 (solid lines) converging more slowly than _ϵ_=0.05 (dotted lines).

Figure 3

Figure 3

Transmission chains for categories. (a) If we consider categories that are sets of four objects defined on three binary dimensions and ignore the assignment of the dimensions to the physical properties of those objects, there are just six possible category types (i–vi), types I–VI (Shepard et al. 1961). Each of the six types is illustrated on a cube, where each dimension of the cube corresponds to one of the binary dimensions and the vertices are the eight objects. Filled circles represent members of an example category of that type. Type I categories are defined on one dimension; type II uses two dimensions; types III, IV and V are one dimension plus an exception and type VI uses all three dimensions. (b) Transmission chains were constructed by showing people three objects drawn from a category and asking them to indicate, from a set of possible alternatives, which object completed the set. The objects seen by the next person were selected at random from the set selected by the previous person. The probability with which people selected categories of the six types changes as a function of the number of generations of a transmission chain, as predicted by a Bayesian model using a prior estimated from human learning data. In particular, the probabilities of types I and VI increase and decrease, respectively. (i) Human participants and (ii) Bayesian model (circles, type I; crosses, type II; triangles, type III; squares, type IV; five-point stars, type V; six-point stars, type VI). Further details are provided in Griffiths et al. (2008).

Figure 4

Figure 4

Representative results for transmission chains with human participants in which people learn functions. (a–d) Each row shows a single chain. (i) The (x, y) pairs were presented to the first participant in the chain, being represented as the width and height of horizontal and vertical rectangles, respectively. Participants then made predictions of the value of y for new x values ((ii) _n_=1, (iii) _n_=2, (iv) _n_=3, (v) _n_=4, (vi) _n_=5, (vii) _n_=6, (viii) _n_=7, (ix) _n_=8, (x) _n_=9). These predictions formed the (x, y) pairs given to the next person in the chain, whose data appear in (ii)–(x) and so forth. Consistent with the previous research exploring human inductive biases for function learning, chains produced linear functions with mostly positive slopes, regardless of whether they were initialized with (a) a positive linear function, (b) a negative linear function, (c) a nonlinear function or (d) a random collection of points.

Figure 5

Figure 5

The interaction of selection with inductive biases. (a) Increasing the selective pressure in favour of hypothesis 1 increases the representation of that hypothesis in the population. The equilibrium probability of hypothesis 1 for _π_=0.2, _ϵ_∈{0.01,0.05} (solid line, dotted line, respectively), and a range of values of the selective pressure s are shown. (b) Threshold on s for hypothesis 1 to obtain an equilibrium probability greater than 0.5 as a function of π and ϵ. For values of π and ϵ such that _q_21>0.5, no value of s produces an equilibrium favouring hypothesis 1.

Similar articles

Cited by

References

    1. Anderson J.R. Erlbaum; Hillsdale, NJ: 1990. The adaptive character of thought.
    1. Anderson J.R. The adaptive nature of human categorization. Psychol. Rev. 1991;98:409–429. doi:10.1037/0033-295X.98.3.409 - DOI
    1. Anderson J.R, Milson R. Human memory: an adaptive perspective. Psychol. Rev. 1989;96:703–719. doi:10.1037/0033-295X.96.4.703 - DOI
    1. Ashby F.G, Alfonso-Reese L.A. Categorization as probability density estimation. J. Math. Psychol. 1995;39:216–233. doi:10.1006/jmps.1995.1021 - DOI
    1. Atran S. The trouble with memes: inferences versus imitation in cultural creation. Hum. Nat. 2001;12:351–381. doi:10.1007/s12110-001-1003-0 - DOI - PubMed

Publication types

MeSH terms

LinkOut - more resources