Is Explanation By Ideal Types A Good Model For The Social Sciences?

In what follows I will argue that Weber’s ideal type model of explanation is inadequate for the social sciences.  I will outline the theory of explanation by ideal types.  I’ll then show it to be inadequate because the ideal type model falsely assumes that the ideally rational agent acts from only a single consideration, for example an economic consideration.

Weber contended that one could causally explain social phenomena by employing a comparative method looking at ideal types.  The social scientist first straightens out in his own mind how his subject would have acted if he had been entirely rational in the pursuit of his ends.  Then the social scientist compares this idealised model with the real actions of the subject.  Finally the social scientist gives a causal explanation of any observed deviation of his subject’s actions from those dictated by the ideal type model.  Ideal types are analytical conceptual constructs that abstract certain features of social phenomena3.  They are clear and internally coherent, which Weber likes because it makes his methodology objective and scientific despite the social scientist having to place himself in the mind of his subject4.  Weber distinguishes four ideal types: zweckrational, wertrational, affective, traditional.  The traditional ideal type expresses a settled custom and the affective ideal type accounts for the subject’s unconscious emotions.  The wertrational type embodies value-rational actions.  The agent systematically organises his actions to achieve a certain important value.  This is rational, according to Weber, because the agent consciously decides on his ultimate goal (value) and acts to achieve it.  In this sense traditional and affective ideal types are not rational because, respectively, the goal is ingrained a priori in the subject or the subject does not consciously select the value to which he acts1.  Wertrational actions are not perfectly rational, as Weber admits, as they concentrate on ends rather than the means of achieving said end1.  Zweckrational actions are rational.  The zweckrational ideal type embodies instrumental rationality, that is, the agent employs the best means to achieve a given end.  This type is implicit in orthodox microeconomics and idealised in expected utility theory2.  As this is the most ideal (rational) ideal type, I will use this type to criticise Weber’s methodology.

Zweckrational actions are ideal because they abstract from every consideration that is not an economic consideration and because they abstract to an ideally rational agent2.  Now clearly, these two assumptions essentially abstract to a situation wholly different from the real world in which people are not perfectly rational and have several simultaneous concerns.  One might question how Weber’s theory can answer real world questions when he starts from these false assumptions.  If we take a very basic example drawn from economics we can see that Weber intends the ideal types to be yard sticks by which to measure deviations.  The social scientist may reconstruct an expected utility calculation.  He may ask why was Joe’s choice of oranges instead of lemons rational given his preferences, information, and resources.  To understand the social scientist must identify the ideal solution to Joe’s problem of choice and then apply it as a yardstick to highlight and measure deviations.  So, if Joe chose oranges because he enjoyed eating an orange to a sour lemon and thus derived a higher utility from purchasing the oranges, this would be a rational choice and the reconstruction would tell the social scientist how Joe arrived at it.  If, on the other hand, Joe chose the oranges over the lemons because he always did, that is, for no other reason than tradition then Joe did not make a rational choice and the reconstruction would tell the social scientist what needs further explanation, that is, Joe’s failure to act rationally2.  Hence, Weber is arguing that in order to understand and explain why Joe bought oranges over lemons it is first necessary to work out whether that was Joe’s best decision; the best means to achieve his given ends (in this example the rational choice and game theorist assumes his end is the highest possible utility).  To do this, I agree with Weber that the social scientist needs to abstract to an ideally rational decision-maker because the best decision would be rational (indeed often we assume when observing a stranger’s behaviour that they are acting in that way because they believe it to be the best possible means of attaining their goal, I think this assumption comes from the individual knowing he acts in the way he perceive best to achieve his end and extrapolates this to others).

However, I disagree with Weber that the social scientist needs to abstract to a world in which Joe’s only concern is his economic transaction.  On the contrary, I think the social scientist has to take all Joe’s considerations into account to determine Joe’s most rational course of action.  Weber, of course concedes that the ideal type model of social action constructed with the false premise that Joe would be acting in ideally rational pursuit of purely economic goals is unrealistic1.  But he argues that ideal type methodology enables social scientists to understand people’s real actions (even when their actions are shaped by customs, emotional impulses, errors, and the influence of non-economic concerns) because, and to the extent that, people are also affected by the rational pursuit of economic ends either in particular cases or on average1.  Similarly, the ideal type model facilitates knowledge of men’s real motives by making use of the very deviation of the actual course of events from the ideal type1.  It seems to me that Weber is arguing that in the real world most people act in a confused state of unconsciousness only dimly aware of their intentions.  They may act form impulse or custom.  But the ideal type model is still preferable for social science because it is more scientific.  It is more scientific because it can be separated out and made explicitly clear.  The ideal type model is then the least worst.  But this account seems to miss things out.  What if Joe chose oranges over lemons because he has a wholly irrational fear of the colour yellow, then the ideal type model might explain this in terms of disutility from the purchase of lemons and so it is rational to chose oranges.  But the disutility from the lemons is irrational, so is not Joe’s choice irrational also?  Or what if Joe has an irrational fear of lemons but buys them anyway because his ailing mother incorrectly (as Joe knows) believes they will cure her insomnia, then what would the ideal types advocate conclude?  In seeking clarity and neatness Weber seems to loose explanatory power.  In this last example Joe’s intended meaning is clear in his own mind and his actions are rational in the sense that they focused on achieving his given end but ideal type models would be useless to the social scientist because Joe is operating outside the realms of economics.  Weber seems to take rational choice and game theories as gold standard, but it could be that these theories are inadequate to explain accurately the actions of social actors all the time.  Specifically they are inadequate to explain the agent’s actions when he is acting out of more than simply economic considerations.

Further, in that last example, Joe is operating outside the realm of himself, that is, he is not acting as an individual anymore he is taking his mother and her beliefs into consideration.  Weber methodology requires him to be individualist.  He must start with individual actors so that he can identify two different types of meaning of action.  Subjective meaning asks what the actor meant by his action and inter-subjective meaning asks what the action meant2.  This is so he can concentrate on subjective meaning in his methodology and forget what the action meant to focus on intended meaning.  This implies that individual intensions are prior the shared reading that formulates society.  One could argue that intensions are only possible in light of public ‘rules of the game’2.  Perhaps Weber is right to adhere to rational choice theory and game theory when looking at an individual with only economic considerations to suggest a method of reconstructing the agent’s calculation, but this does not translate to an individual in society.  For the individual in society – having mixed considerations and intensions, and being necessary to cooperate and maintain relationships (and perhaps valuing certain relations not for the ends they might help him to obtain, indeed they may obscure these ends, but for their intrinsic good), also where the individual’s ends are inconstant and sometimes inconsistent – it may well be rational to follow a rule.  And Weber’s ideal type model does not account for this.

In sum, Weber’s ideal type model is, overall, an inadequate model of explanation for the social sciences.  He correctly identifies that in order to understand and explain an agent’s behaviour one should abstract to an ideally rational agent and explain any deviations between the ideal and the real agent’s actions.  This is correct because we already assume that an agent will always try to follow the most rational path, that is, the best means to achieve their given end.  Perhaps this assumption comes from each member of society looking inwardly and identifying this as his own way of acting.  In asking why an individual did not act as an ideally rational agent would, Weber can discover new reasons for social actions.  However, this is where the ideal type model falls down.  The assumption that the best course of action is the rational course of action, and my assumption that this is the way people generally try to behave, means that Weber cannot abstract to an agent whose only consideration is, for instance, economic.  This is an oversimplification.  The real world agent may well be acting perfectly rationally but game theory or rational choice theory would not pick this up because the agent is acting with several (perhaps contradictory) considerations in mind and he is could be acting outside of himself, that is, he is taking into account the considerations of others.  Further, when focusing on the individual Weber may not account for rational behaviour, which is rational not for sparking the most utility for the agent but because such behaviour follows the rules of society (the rules of the game).  This behaviour may seem irrational when looking only at the individual because it might deny him the ends he has chosen for himself.



1.Weber, M (1978).  Chapter 1 – The Nature of Social Actions.  Economy and Society: an interpretive sociology.  Berkeley University of California Press.

2.Hollis, M (2002).  The Philosophy of Social Science an Introduction.  Cambridge University Press.

3.Blackburn, S (2005).  Oxford Dictionary of Philosophy.  Oxford University Press.

4.Honderich, T (2005).  Oxford Encyclopaedia of Philosophy.  Oxford University Press