Mechanistic explanation at the limit

Resurgent interest in both mechanistic and counterfactual theories of explanation has led to a fair amount of discussion regarding the relative merits of these two approaches. James Woodward is currently the pre-eminent counterfactual theorist, and he criticizes the mechanists on the following grounds: Unless mechanists about explanation invoke counterfactuals, they cannot make sense of claims about causal interactions between mechanism parts or of causal explanations put forward absent knowledge of productive mechanisms. He claims that these shortfalls can be offset if mechanists will just borrow key tenets of his counterfactual theory of causal claims. What mechanists must bear in mind, however, is that by pursuing this course they risk both the assimilation of the mechanistic theories of explanation into Woodward’s own favored counterfactual theory, and they risk the marginalization of mechanistic explanations to a proper subset of all explanations. An outcome more favorable to mechanists might be had by pursuing an actualist-mechanist theory of the contents of causal claims. While it may not seem obvious at first blush that such an approach is workable, even in principle, recent empirical research into causal perception, causal belief, and mechanical reasoning provides some grounds for optimism.     

Moving Beyond Causes: Optimality Models and Scientific Explanation

A prominent approach to scientific explanation and modeling claims that for a model to provide an explanation it must accurately represent at least some of the actual causes in the event's causal history. In this paper, I argue that many optimality explanations present a serious challenge to this causal approach. I contend that many optimality models provide highly idealized equilibrium explanations that do not accurately represent the causes of their target system(s). Furthermore, in many contexts, it is in virtue of their independence of causes that optimality models are able to provide a better explanation than competing causal models. Consequently, our account of explanation and modeling must expand beyond the causal approach.     

Information becomes evidence when an explanation can incorporate it into a causal framework

Recognizing information as evidence is central to the development of scientific reasoning. When does information about an event come to be treated as evidence relevant to explaining the event? We asked whether this was increasingly likely to happen when an explanation becomes available that can incorporate both the event and the information into a single causal framework. In three studies, we presented participants with events for which there were two possible and plausible explanations (a baseline and one of two alternative explanations), as well as with two pieces of background information. While all explanations could account for the event, only one alternative explanation (the “target” explanation) could incorporate both the event and the background information into a single causal framework. The results indicated that information is more likely to be seen as evidentially relevant to an event when there is an explanation available that can accommodate both the event and the information into a single casual framework than when such an explanation is lacking. Furthermore, the presence of this information renders the target alternative increasingly plausible. That is, it is the interdependence of explanation or theory and evidence that allows us to realize that some information is likely to be evidential. However, for this to happen, the relation between explanation and information must be made salient, either by explicitly asking about it (as we did in Study 1) or by fleshing out the target explanation (as we did in Study 3).     

An ideal solution to disputes about multiply realized kinds

Multiply realizable kinds are scientifically problematic, for it appears that we should not expect discoveries about them to hold of other members of that kind. As such, it looks like MR kinds should have no place in the ontology of the special sciences. Many resist this conclusion, however, because we lack a positive account of the role that certain realization-unrestricted terms play in special science explanations. I argue that many such terms actually pick out idealizing models. Idealizing explanation has many of the features normally associated with explanation by MR kinds. As idealized models are usually mere possibilia, such explanations do not run afoul of the metaphysical problems that plague MR kinds.

Bayesian Fundamentalism or Enlightenment? On the explanatory status and theoretical contributions of Bayesian models of cognition

The prominence of Bayesian modeling of cognition has increased recently largely because of mathematical advances in specifying and deriving predictions from complex probabilistic models. Much of this research aims to demonstrate that cognitive behavior can be explained from rational principles alone, without recourse to psychological or neurological processes and representations. We note commonalities between this rational approach and other movements in psychology - namely, Behaviorism and evolutionary psychology - that set aside mechanistic explanations or make use of optimality assumptions. Through these comparisons, we identify a number of challenges that limit the rational program's potential contribution to psychological theory. Specifically, rational Bayesian models are significantly unconstrained, both because they are uninformed by a wide range of process-level data and because their assumptions about the environment are generally not grounded in empirical measurement. The psychological implications of most Bayesian models are also unclear. Bayesian inference itself is conceptually trivial, but strong assumptions are often embedded in the hypothesis sets and the approximation algorithms used to derive model predictions, without a clear delineation between psychological commitments and implementational details. Comparing multiple Bayesian models of the same task is rare, as is the realization that many Bayesian models recapitulate existing (mechanistic level) theories. Despite the expressive power of current Bayesian models, we argue they must be developed in conjunction with mechanistic considerations to offer substantive explanations of cognition. We lay out several means for such an integration, which take into account the representations on which Bayesian inference operates, as well as the algorithms and heuristics that carry it out. We argue this unification will better facilitate lasting contributions to psychological theory, avoiding the pitfalls that have plagued previous theoretical movements.     

Counterfactuals,Probabilities, and Information: Response to Critics

In earlier work we proposed an account of information grounded in counterfactual conditionals rather than probabilities, and argued that it might serve philosophical needs that more familiar probabilistic alternatives do not. Demir [2008] and Scarantino [2008] criticize the counterfactual approach by contending that its alleged advantages are illusory and that it fails to secure attractive desiderata. In this paper we defend the counterfactual account from these criticisms, and suggest that it remains a useful account of information.     

Mental contrasting of counterfactual fantasies attenuates disappointment,regret, and resentment

Negative emotions elicited by positive counterfactuals about an alternative past—“if only” reconstructions of negative life events—are functional in preparing people to act when opportunities to restore the alternative past will arise. If the counterfactual past is lost, because restorative opportunities are absent, letting go of the negative emotions should be the better solution, sheltering people from feelings of distress. In six experimental studies, the self-regulation strategy of mental contrasting (Oettingen, European Review of Social Psychology 23:1–63, 2012) attenuated the negative emotions elicited by positive fantasies about a lost counterfactual past, specifically, disappointment, regret and resentment. Mental contrasting (vs. relevant control conditions) led people to feel less disappointed when evaluating their lost counterfactual past compared with their current reality, indicating reduced commitment to the lost counterfactual past (Studies 1, 2, 3, and 4), and it attenuated post-decisional regret and resentment (Studies 5 and 6). These findings held when participants were induced to focus on lost counterfactual pasts for which they were responsible (Studies 4 and 5), for which they blamed another person (Study 6), or for which they deemed no one responsible (Studies 2 and 3). The findings are relevant for building interventions that help people to come to terms with their lost counterfactual past.     

Reduction: the Cheshire cat problem and a return to roots

In this paper, I propose two theses, and then examine what the consequences of those theses are for discussions of reduction and emergence. The first thesis is that what have traditionally been seen as robust, reductions of one theory or one branch of science by another more fundamental one are a largely a myth. Although there are such reductions in the physical sciences, they are quite rare, and depend on special requirements. In the biological sciences, these prima facie sweeping reductions fade away, like the body of the famous Cheshire cat, leaving only a smile. ... The second thesis is that the “smiles” are fragmentary patchy explanations, and though patchy and fragmentary, they are very important, potentially Nobel-prize winning advances. To get the best grasp of these “smiles,” I want to argue that, we need to return to the roots of discussions and analyses of scientific explanation more generally, and not focus mainly on reduction models, though three conditions based on earlier reduction models are retained in the present analysis. I briefly review the scientific explanation literature as it relates to reduction, and then offer my account of explanation. The account of scientific explanation I present is one I have discussed before, but in this paper I try to simplify it, and characterize it as involving field elements (FE) and a preferred causal model system (PCMS) abbreviated as FE and PCMS. In an important sense, this FE and PCMS analysis locates an “explanation” in a typical scientific research article. This FE and PCMS account is illustrated using a recent set of neurogenetic papers on two kinds of worm foraging behaviors: solitary and social feeding. One of the preferred model systems from a 2002 Nature article in this set is used to exemplify the FE and PCMS analysis, which is shown to have both reductive and nonreductive aspects. The paper closes with a brief discussion of how this FE and PCMS approach differs from and is congruent with Bickle’s “ruthless reductionism” and the recently revived mechanistic philosophy of science of Machamer, Darden, and Craver.     

Natural Selection: A Case for the Counterfactual Approach

This paper investigates the conception of causation required in order to make sense of natural selection as a causal explanation of changes in traits or allele frequencies. It claims that under a counterfactual account of causation, natural selection is constituted by the causal relevance of traits and alleles to the variation in traits and alleles frequencies. The “statisticalist” view of selection (Walsh, Matthen, Ariew, Lewens) has shown that natural selection is not a cause superadded to the causal interactions between individual organisms. It also claimed that the only causation at work is those aggregated individual interactions, natural selection being only predictive and explanatory, but it is implicitly committed to a process-view of causation. I formulate a counterfactual construal of the causal statements underlying selectionist explanations, and show that they hold because of the reference they make to ecological reliable factors. Considering case studies, I argue that this counterfactual view of causal relevance proper to natural selection captures more salient features of evolutionary explanations than the statisticalist view, and especially makes sense of the difference between selection and drift. I eventually establish equivalence between causal relevance of traits and natural selection itself as a cause.     

A model-based comparison of three theories of audiovisual temporal recalibration

Observers change their audio-visual timing judgements after exposure to asynchronous audiovisual signals. The mechanism underlying this temporal recalibration is currently debated. Three broad explanations have been suggested. According to the first, the time it takes for sensory signals to propagate through the brain has changed. The second explanation suggests that decisional criteria used to interpret signal timing have changed, but not time perception itself. A final possibility is that a population of neurones collectively encode relative times, and that exposure to a repeated timing relationship alters the balance of responses in this population. Here, we simplified each of these explanations to its core features in order to produce three corresponding six-parameter models, which generate contrasting patterns of predictions about how simultaneity judgements should vary across four adaptation conditions: No adaptation, synchronous adaptation, and auditory leading/lagging adaptation. We tested model predictions by fitting data from all four conditions simultaneously, in order to assess which model/explanation best described the complete pattern of results. The latency-shift and criterion-change models were better able to explain results for our sample as a whole. The population-code model did, however, account for improved performance following adaptation to a synchronous adapter, and best described the results of a subset of observers who reported least instances of synchrony.     

The explanatory potential of artificial societies

It is often claimed that artificial society simulations contribute to the explanation of social phenomena. At the hand of a particular example, this paper argues that artificial societies often cannot provide full explanations, because their models are not or cannot be validated. Despite that, many feel that such simulations somehow contribute to our understanding. This paper tries to clarify this intuition by investigating whether artificial societies provide potential explanations. It is shown that these potential explanations, if they contribute to our understanding, considerably differ from potential causal explanations. Instead of possible causal histories, simulations offer possible functional analyses of the explanandum. The paper discusses how these two kinds explanatory strategies differ, and how potential functional explanations can be appraised.     

Negation and its impact on the accessibility of text information

Prior experiments have shown that sentences such as (1) Mary bakes bread but no cookies lead to a reduced accessibility of the concept mentioned in the negated phrase, whereas sentences such as (2) Elizabeth burns the letters but not the photographs do not. In the present article, two explanations for this result are investigated. According to situation model theory (Johnson-Laird, 1983; van Dijk & Kintsch, 1983), the reason is that the entity mentioned within the negated phrase in (2) is not absent from the described situation. According to discourse representation theory (Kamp, 1981), in contrast, the negation operator in (2) does not reduce the accessibility of the negated concept, because the corresponding discourse referent is not introduced but merely referred to within the operator's scope. In two experiments, participants were presented with narrative texts including negation sentences that either introduced or referred to entities, and that either described a situation in which only the nonnegated or only the negated entity was present. The accessibility of the relevant concepts was measured by means of a probe recognition task. The results support the situation models explanation.     

Folk psychology as science

There is a long-standing debate in the philosophy of action and the philosophy of science over folk psychological explanations of human action: do the (perhaps implicit) generalizations that underwrite such explanations purport to state contingent, empirically established connections between beliefs, desires, and actions, or do such generalizations serve rather to define, at least in part, what it is to have a belief or desire, or perform an action? This question has proven important because of certain traditional assumptions made about the role of law-statements in scientific explanations. According to this tradition, law-statements take the form of generalizations, and the laws we find in well-established sciences are contingent and empirical; as such, if the kinds of generalizations at work in folk psychological explanations of human action act like definitions, or state conceptual connections, then such generalizations could not play the kind of explanatory role we find in mature sciences. This paper argues that the aforementioned way of framing the debate reflects a still powerful but impoverished conception of the role laws play in scientific explanations, a conception that, moreover, cannot be reconciled with a good deal of actual scientific practice. When we update the philosophy of science, we find the concerns that are raised for folk psychological explanations largely evaporate or are found not to be specific to such explanations.     

Generating explanations via analogical comparison

Generating explanations can be highly effective in promoting learning in both adults and children. Our interest is in the mechanisms that underlie this effect and in whether and how they operate in early learning. In adult reasoning, explanation may call on many subprocesses—including comparison, counterfactual reasoning, and reasoning by exclusion; but it is unlikely that all these processes are available to young children. We propose that one process that may serve both children and adults is comparison. In this study, we asked whether children would use the results of a comparison experience when asked to explain why a model skyscraper was stable. We focused on a challenging principle—that diagonal cross-bracing lends stability to physical structures (Gentner et al., Cognitive Science, 40, 224–240, 2016). Six-year-olds either received no training or interacted with model skyscrapers in one of three different conditions, designed to vary in their potential to invite and support comparison. In the Single Model condition, children interacted with a single braced model. In the comparison conditions (Low Alignability and High Alignability), children compared braced and unbraced models. Following experience with the models, children were asked to explain why the braced model was stable. They then received two transfer tasks. We found that children who received highly alignable pairs were most likely to (a) produce brace-based explanations and (b) transfer the brace principle to a dissimilar context. This provides evidence that children can benefit from analogical comparison in generating explanations and also suggests limitations on this ability.     

Causal Explanation and Explanatory Selection

It is observed that in ordinary everyday causal explanations often just one causal factor is mentioned. One causal factor carries the explanatory burden, even if there are several causal factors that are responsible for the event to be explained. This paper deals with the question of how to account for this explanatory selection. I argue for a pragmatic stance towards explanation, that we must attend to the question–answer situation as a whole and the context of the explanation. The context of an explanation includes the inquirer's and the explainer's beliefs and presuppositions relevant for the explanation-seeking question, and these are encoded in a reference class. Furthermore I argue that the explanation-giving answer contains an implicit counterfactual claim, the explanation-giving counterfactual. The solution to the problem of explanatory selection is to be found in the presuppositions encoded by the reference class and the eg-counterfactual. In short we select as explanatory that factor which, together with the presupposition encoded in the reference class we believe will make the eg-counterfactual true. This revised version was published online in June 2006 with corrections to the Cover Date.     

Implicit bias,confabulation, and epistemic innocence

In this paper I explore the nature of confabulatory explanations of action guided by implicit bias. I claim that such explanations can have significant epistemic benefits in spite of their obvious epistemic costs, and that such benefits are not otherwise obtainable by the subject at the time at which the explanation is offered. I start by outlining the kinds of cases I have in mind, before characterising the phenomenon of confabulation by focusing on a few common features. Then I introduce the notion of epistemic innocence to capture the epistemic status of those cognitions which have both obvious epistemic faults and some significant epistemic benefit. A cognition is epistemically innocent if it delivers some epistemic benefit to the subject which would not be attainable otherwise because alternative (less epistemically faulty) cognitions that could deliver the same benefit are unavailable to the subject at that time. I ask whether confabulatory explanations of actions guided by implicit bias have epistemic benefits and whether there are genuine alternatives to forming a confabulatory explanation in the circumstances in which subjects confabulate. On the basis of my analysis of confabulatory explanations of actions guided by implicit bias, I argue that such explanations have the potential for epistemic innocence. I conclude that epistemic evaluation of confabulatory explanations of action guided by implicit bias ought to tell a richer story, one which takes into account the context in which the explanation occurs.     

The effect of explanation in simple binary decision tasks

Many studies of explanation have focused on higher level tasks and on how explanations draw upon relevant prior knowledge, which then helps in understanding some event or observation. However, explanations may also affect performance in simple tasks even when they include no task-relevant information. In three experiments, we show that explanations adding no task-relevant information alter performance in a sequential binary decision task. Whereas people with no explanation for why two events occurred at different rates tended to predict each outcome in proportion to their probability of occurrence (to "probability match"), people with an explanation tended to predict the more likely event more often (to "overmatch," a better strategy). These results suggest a broader view of explanation, which includes a role in shaping simple tasks outside of higher level reasoning.