Explanations in Social Science I ‘The main task of the social sciences is to explain social phenomena. It is not the only task, but it is the most important one, to which others are subordinated or which they depend’ (Elster 2007: 20). Outline of session 1. Explanations in social science 2. The ‘Covering Law’ model of Explanation 3. Causal explanation 4. An introduction to functionalism and structuralism 1. Explanations in social science The Hollis Matrix Understanding Explanation Holism Systems 'Games’ Atomism Agents Actors 4 The Hollis Matrix: Explanation Understanding Explanation Holism Functionalism / Structuralism Atomism Rational Choice / Game Theory 5 This week: Holist Explanation Understanding Explanation Holism Functionalism / Structuralism Atomism 6 Types of explanation General Structure: Explanans > Explanandum Types of Explanation • • • • event > fact > event > fact > event explanations event explanations fact explanations fact explanations Types of explanation • event > event explanations (the 2001 Gulf War is explained by Iraq’s invasion of Kuwait) • fact > event explanations (the slippery road caused the motorway crash) • event > fact explanations (the motorway crash explains why many people did not work on Monday) • fact > fact explanations (record low mortgage rates in the UK in 2014 explains why average house prices rose 10% in 2014) Supporting an explanation 1. …from below (‘excess explanatory power’) 2. …from above (flows from more general theory) 3. …lateral support (rival explanations flawed) ‘In the long run it is the theory that is supported by the successful explanations it generates, not the other way around (Elster 2007: 20). 2. The ‘Covering Law’ model of explanation • Scientific explanation as “the idea that a given event or regularity can be subsumed under one or more general laws” (Little: 3). • A good explanation shows that a phenomenon (the explanandum) follows from a combination of law-like generalizations and facts (the explanans). • Works for particular facts/events (dead plant in dark study) and processes (photosynthesis) which are deduced from still more general (chemical and physical) laws. • Three conditions for a good scientific explanation (Hempel 1965): 1. the premises entail the conclusion (deductively valid) 2. the premises are true 3. the premises contain at least one general law [L]. 1+2+3= The ‘Covering Law Model’ of explanation 2. The ‘Covering Law’ model (cont.) P1: Plants survive only if they can photosynthesize [L]. P2: Plants cannot photosynthesize without sunlight [L]. P3: Ed’s office is so dark that his plants do not receive any sunlight [F]. P3: Ed’s office plants cannot photosynthesize [L]. C: Ed’s office plants are all dead. Critique 1. Questioning symmetry 2. Explanation v irrelevant truth 3. Explanation v necessity (1) Questioning symmetry • P1: light hits the flagpole which is 15m high (F) • P2: angle of elevation is 37o (F) • P3: light travels in straight lines (L) • P4: laws of trigonometry (L) • C: flagpole will cast a shadow of 20m (tan 37o=15/20) (Okasha 2002) It is true that we can deduce/predict the height of the flagpole from the length of the shadow but this does not explain why the flagpole has the height it does! (2) Explanation v irrelevant truth In the maternity ward… P1: John is taking birth control pills (F) P2: People who take birth control pills do not get pregnant (L) C: John is not pregnant This is not a good explanation, but it fits the covering law model. A good explanation must contain information that is relevant to the occurrence of the phenomenon being explained! (3) Questioning necessity The Sherlock Holmes Objection… (3) Questioning necessity P1: Watson had chalk between his left finger and thumb when he returned from the club last night [F] P2: Watson puts chalk there when he plays billiards to steady the cue [L] P3: Watson never plays billiards except with Thurston [F] P4: Watson told Holmes 4 weeks ago that Thurston had an option on some South African property that would expire in a month, and which he desired Watson to share with him[F]. P5: Watson’s chequebook is locked in Holmes’s drawer and Watson has not asked for the key [F]. C1: Watson does not propose to invest his money in South African securities [Holmes’s preferred explanation] C2: Watson proposes to invest his money in South African Securities without using his chequebook [rival explanation equally well confirmed by data]. [see Arthur Conan Doyle, ‘The Dancing Man’] 3. Causal explanation • When establishing a causal relationship such that event/fact x causes event/fact y (x→y), x must be a ‘substantial factor’ in y coming about. • What is this ‘substantial factor’? • Common approach: y would not have occurred ‘but for’ x (causein-fact) • It is not necessary for establishing causation that the presence of x guarantees the occurrence of y (sufficient causation) • Neither is it necessary that it was foreseeable (e.g. by the agent) that x would bring about y (proximate cause) - though compare with tort law • Sometimes there may be multiple prior conditions (x1,x2,..xn) separately capable of bringing about y (causal over-determination) • Social scientists are interested in identifying instigating causes (new causes in fact) as opposed to standing causes (permanent causes in fact) Causation in natural and social science: four tests • What is causal explanation?→y), four questions arise (Kellstedt and Whitten 2008: Ch.4). • Any causal explanation should be put through these four ‘hurdles’ before it is established as a valid scientific explanation that x causes, and thus explains the occurrence of, y. • An ideal causal explanation of x→y will result in a the pattern of answers: • 1:yes, 2:no, 3:yes, 4: no • Any deviation from this pattern will a potential source of critique/doubt as to the proposed causal relation. 1. Is there a credible mechanism underpinning the x→y causal claim? [how are babies and shadows produced; answer may be probabilistic]. 2. Are there good reasons for thinking, on the contrary, that y causes x (the assumption being that x cannot be the cause of y if y is the cause of x)? [what is the direction of causality between flagpoles and shadow] 3. Is there a measurable association between the occurrence of x and y (ie do x and y co-vary)? [does the frequency/size of flagpoles rise and/or fall with that of shadows] 4. Is there a further, connected factor, z, rendering the observed relationship between x and y spurious? [does another factor, biology, explain why John is not pregnant and hence the co-variance of his birth control and his non-pregnancy] Causal explanation (cont.) (a) Flagpole height→Shadow length: 1: yes, 2: no, 3: yes, 4: no. (b) Shadow length→Flagpole height: 1: no, 2: yes, 3: yes, 4: no. (c) John’s tablets→John’s non-pregnancy: 1: no, 2: no, 3:yes, 4:yes. 4. Functionalist and structuralist explanation Functionalism: attributes purpose at the macro level and from there draws inferences about the micro level Structuralism: infers macro level outcomes from the influence of social structures upon individual behaviour Functionalism Particular social phenomena explained in terms of the function they perform in the general social system. Eg Durkheim: crime, religion, family, democracy (‘social facts’) can be explained by their positive role in securing social stability and the ‘systemic normalcy’ and ‘health’ (v pathology) of the social system Model of explanation • An institution, background pattern, or trait [T] persists in Society [S] • Why does this trait [T] persist? A. [T] has certain effects [E] B. These effects are functional/adaptive: their presence solves the problem of designing a trait to assure that the needs of the society are satisfied C. [T] persists in S because it has the disposition to produce [E]. Example 1: Durkheim Why does crime [T] persist? A. Punishing crime has effects including the entrenchment of the values that crime offends B. These effects are functional/adaptive: entrenching the values that crime offends contributes social stability C. T persists in society because it has the disposition to produce social stability. Durkheimian concepts: ‘social facts’, ‘systemic normality’, ‘pathology’, ‘health’, ‘manifest function’, ‘latent function’ Example 2: structural-functionalism Why do the ceremonial weeping rites of the Andaman islands [T] persist? A. Burial and mourning rituals of the islands have predictable effects including the entrenchment of a collective cultural identity B. These effects are functional/adaptive: entrenching a collective identity contributes to social stability C. T persists in society because it has the disposition to produce social stability [Radcliffe Brown (1951) / Little (1991: 94-5)]. ‘To explain a social fact it is not enough to show the cause on which it depends; we must also, at least in most cases, show its function in the establishment of social order’ (Durkheim 1950: 97). Critique • ‘The mere fact an activity has beneficial consequences is not sufficient to explain it’ (Elster 1986). • How does a society arrange its institutions to meet its needs (Little 1991)? • Ignores individual-level intentions and moral agency (not a problem in biology) • The “lamppost” approach: is looking for any available answer, even if it is not the one we need, better than peering into the dark • Why THIS and not ANOTHER activity? (E.g. burial rites). Structuralism ‘[Structural explanations] explain a feature of society as the predictable consequence of certain structural characteristics of society’ (Little 1991:91) Structuralism Structuralism: existence of particular social phenomena (social norms, beliefs, social institutions) only understandable as necessary elements of a larger social structure. (1) social structures are prominent causes of social phenomena (causal structuralism) and (2) social phenomena are explained by how they fit into abstract underlying structures (noncausal structuralism). Model of explanation 1: Causal 2. Non-causal 1. The existence of a particular social phenomena (social unrest, war) 2. Is brought about by underlying structural features of society (enduring, independent & constraining regulative systems: wagelabour capitalism, alliance structures, bureaucratic norms) 3. Which shape the environment of choice of individual action 1. The existence of a particular social phenomena (social unrest, war) 2. Is explained by an underlying ‘grammar’ of abstract underlying social structures 3. Because these phenomena ‘fit’ ‘underlying structures that possess an abstract order akin to a syntax in language that can be decoded’ (Little 107-8) Example: historical materialism 1. societies are complex systems made up of social structures (eg economic systems) and social phenomena (eg forms of government) can generally be explained as the causal consequences of the particular details of underlying structures. a. The level of development of productive power explains the nature of the economic structure (the sum of productive relations) b. The economic structure explains the nature of the superstructure c. The superstructure stabilizes the economic structure d. The superstructure can be explained by the way it advances productive power (See Cohen 1988). “The mode of production of material life conditions the social, political and intellectual life process in general. It is not the consciousness of men that determine their being but, on the contrary, their social being that determines their consciousness” (Marx 1859). Critique • If it’s causal….over-reliant on standing causes (Little 2002: 104)? But what are the instigating causes of social change? • If it’s non-causal…what’s the added value of knowing the ‘grammar’ when the puzzling similarities/dissimilarities between societies remain unexplained? • The “purification” exercise distorts social reality. It is rarely possible to disentangle social phenomena from each other sufficiently to establish neat variables. Further Rading • Cohen, G.A. (1988) History, Labour and Freedom (Oxford: OUP). • Durkheim, (1950[1895]) The Rules of Sociological Method. Translated by S. A. Solovay and J. H. Mueller. New York: The Free Press. • Elster, J. (1986) An Introduction to Karl Marx (Cambridge: CUP). • Elster, J. (1989) The Cement of Society: Studies in Rationality and Social Change (Cambridge: CUP) • Elster, J. (2007) Explaining Social Behaviour: More Nuts and Bolts of the Social Sciences(Cambridge: CUP). • Hollis, M. (1994) The Philosophy of Social Science (Colorado: Westview). • Kellstadt and Whitten (2008), ‘Evaluating causal relationships’, in The Fundamentals of Political Science Research (Cambridge: Cambridge University Press), pp.45-66. • Little, D. (1991) Varieties of Social Explanation (Boulder: Westview). • Okasha, S. (2002) Philosophy of Science (Oxford: OUP). • Radcliffe-Brown, A.R. (1952) Structure and Function in Primitive Society (New York: Free Press), pp.133-211. • Root, M. (1993) Philosophy of Social Science (Oxford: Blackwell). • Rosenberg, A. (1995) Philosophy of Social Science (Boulder: Westview).