




已阅读5页,还剩9页未读, 继续免费阅读
版权说明:本文档由用户提供并上传,收益归属内容提供方,若内容存在侵权,请进行举报或认领
文档简介
explanationphilip kitcherphilosophical reflections about explanation are common in the history of philosophy, philosophical reflections and important proposals were made by aristotle, hume, kant and mill. but the subject came of age in the twentieth century with the provision of detailed models of scientific explanation, prominently the covering-law model, which takes explanations to be arguments in which a law of nature plays an essential role among the premises. in the heyday of logical empiricism, philosophers achieved a consensus on the covering law model, but, during the 1960s and 1970s, that consensus was challenged through the recognition of four major kinds of difficulty: first, a problem about the relation between idealized arguments and the actual practice of explaining; second, the difficult of characterizing the underlying notion of a law of nature; third, troubles in accounting for the asymmetries of explanation; and, four, recalcitrant problems in treating statistical explanations.appreciation of these difficulties has led to the widespread abandonment of the covering-law, and currently there is no consensus on how to understand explanation. the main contemporary view seeks to characterize explanation in terms of causation, that is, explanations are accounts that trace the causes of events(states, conditions) explained. other philosopher believes that there is no general account of explanation, and offer pragmatic theories, a third option sees explanation as consisting in the unification of the phenomena. all of these approaches have associated successes, and face particular anomalies.although the general character of explanation is now a subject for philosophical debate, some particular kinds of explanation seems to be relatively well understood. in particular, functional explanations in biology, which logical empiricists found puzzling, now appear to be treated quite naturally by supposing them to make tacit reference to natural selection.1 early historythinking about explanation goes back at least to aristotle, whose discussion of four causation in the posterior analytics can properly be viewed as distinguishing modes of scientific explanation(see aristotle&9) in the modern period, the writings of hume, kant and mill offer many insights on causation, laws and regularities in nature that, sometimes explicitly, sometimes implicitly, propose doctrines about the character of scientific explanation. however questions about scientific explanation became sharply focused in the mid-twentieth century, with the emergence of an orthodoxy about scientific explanation, which, despite its later demise, stands as one of the most significant achievements of the movement known as logical empiricism. the writing of karl popper r.b. braithwaite, emest nagel and especially c.g.hempel, articulate an influential conception. namely, scientific explanations are viewed as arguments in which a statement describing the fact to be explained is derived from premises, at least one of which is a law of nature. the underlying idea is that scientific explanations provide understanding by showing that the phenomena to be explained should be expected as a consequence of the general laws of nature.2 the covering-law modelone important and much-discussed species of scientific explanation according with this general conception is deductive-nomological explanation . in cases of this type, the argument is deductive, consequence of the premises advanced in giving the explanation. d-n explanations may be provided for explananda that describe particular facts or for explananda that announce general regularities. in the former case, there is a simple schema which exhibits the form of the explanationwhere the statement e is the explanandum, describe the fact to be explained, the statements l are laws of nature, and the statements c describe particular facts. it is not hard to construct arguments that accord with this schema and which seem to explain their conclusions: derivations in classical newtonian dynamics that deduce the trajectories of bodies from force laws and initial conditions supply many examples.not all explanatory arguments are deductive. the logical empiricist orthodoxy admitted inductive-statistical explanations as well as d-n explanations, the explanandum is inferred inductively from premises at least one of which is a probabilistic law, for example a statement that assigns a value to the probability with which a particular trait is found among members of a specific class. thus, to cite a famous example of hempels, we may explain why a child, henritta, contracted measles, by nothing that she has been in contract with another child, henry, who has measles, and that a large percentage of children who come into contact with measles patients subsequently come down with measles. imitating the schema for d-n explanation, we can present this modest derivation as follows: henrietta has been in contact with henry, and henry has measles. the frequency with which children in contact with measles patients subsequently acquire measles is 99 percent.here, the role indicates that the inference from premises to conclusion is inductively strong, rather than deductively valid; the figure in brackets reveals the strength of the inference. i-s explanations have to meet several requirements first, the numerical strength of the inductive inference must be high. second, the explanations must meet a requirement of maximal specificity: there must not be known further premises which, if added to the explanans, would change the strength of the inference-as, for example, the inductive reasoning would be modified if we knew that henrietta had received a measles shot, and that children given such shots have a very low probability of acquiring measles. plainly, the explanations that scientists and others actually put forward do not look much like these stripped-down arguments. the logical empiricists claimed only that the everyday provision of explanations could be reconstructed by identifying arguments of d-n or i-s form, and that these reconstructions brought into the open what it was about the explanations that enable them to fulfil their function. in the 1940s and 1950s, many scholars were happy to concede that the convering-law model of explanation, which assimilated explanations to arguments with laws among their premises, worked well as a reconstruction of explanations in the natural sciences especially in physics and chemistry but there were important debates about the application of the model to the social sciences and to explanation in everyday life. controversy focused in particular on the activity of historical explanation. historians construct detailed narratives that appear to explain particular events- the outbreak of the american civil war of henry 8ths discussion of the monasteries. if the covering-law model is correct, then a proper reconstruction of these accounts must expose general laws. are there indeed general laws in history? or are the general laws that underlie historical explanation simply psychological laws that connect motivations of historical actors with their actions?3 four kinds of difficultytroubles with history aside, the covering-law model appeared remarkably successful, a rare example of a convincing solution to a philosophical problem. yet, in the 1960s, it came under sustained attack, and, by the end of the decade, it had been almost entirely abandoned. four separate kinds of consideration contributed to this swift reversal of fortune.first was a complaint, articulated by michael scriven, that perfectly satisfactory explanations can be given, and understood, by people who are quite ignorant of the covering-laws that are essential to supposed reconstruction. it is easy to explain to a friend why there is a mess on the floor by pointing out that your arm knocked the open ink bottle off the desk at which you were writing. perhaps it would be possible for a knowing philosopher of science to cite the general laws that govern the behavior of the bottle and the spilled ink, but this knowledge seems entirely irrelevant to the episode in which the chagrined mess-maker explains what has occurred. at the heart of scrivens complaint lay the recognition that the covering-law model had failed to show how the idealized derivations that supposedly highlight how the explanatory work is done are adapted, in specific local situations, to transmit understanding from one person to another. without a pragmatics of explanations, it was possible to challenge the claim that the structures exposed by the logical empiricists reveal the crucial features that make the explanation successful.a second difficulty resulted from continued inability to provide a satisfactory account of natural laws. from the earliest formulations of the covering-law model, its champions had insisted that not every generalization counts as a law. so-called accidental generalizations cannot discharge any explanatory function: it may be a timeless truth about the universe that all ball games played by a red-haired left-hander who forgoes lunch are won by the opposite team, but that accidental generalization sheds no light on the outcome of any particular game. prior to nelson goodmans formulation of a cluster of difficulties surrounding counterfactuals, induction and laws, the problem of distinguishing laws from accidental generalizations appeared an interesting challenge to the logical empiricist project. once the depth of godmans new riddle of inductions had become apparent, it seemed impossible to find a solution within the constraints that empiricists allowed themselves.a third trouble emerged from recognition that, even if the distinction between laws and accident generalizations could be drawn, the covering-law model would still be too liberal. introducing an example that was to become famous, sylvain bromberger pointed out that the model is blind to certain asymmetries in explanation. we can explain the length of the shadow cast by a flagpole by deriving a statement ascribing the pertinent numerical value from premises identifying the height of the pole and the angle of elevation of the sun, together with law of propagation of the light. this new derivation does not seem explanatory, for it appears wrong to explain the heights of poles in terms of the measurements of the shadows they cast. scrupulous about appealing to causation, logical empiricism had tried to construct an account of explanation without in invoking causal notions that would offend humean sensibilities. brombergers critique suggests that the omission of causal concepts assimilates cases that are importantly different: after all, it is tempting to characterize the difference between the two derivations by pointing out that the height of the flagpole causes the shadow to have the length it does, but that the length of the shadow does not cause the flagpole to have the height it does.perhaps the most influential difficulty, was the fourth, which focused on the failure of the account of statistical explanation. alberto coffa probed the conditions of i-s explanation, revealing that they involved an essential reference to the state of knowledge, which made it impossible to develop a concept of a true inductive explanation. coffas critique complemented the work of richard jeffrey, who had earlier argued that it is possible to explain individual events that do not have high probability in the light of background conditions, and thus that the high probability requirement was also defective. at the same time, wesley salmon worked out, in considerable detail, an account of statistical explanation that, like jeffreys, rejected the thesis that explanations are arguments. central to salmons account was the idea that we explain by citing probabilistically relevant information. in the early versions of his model of statistical explanation, salmon proposed that probabilistically explanations gain their force from the recognition that the probability that an individual has a property has been raised. schematically, the information that a is f helps explain why a is g when the probability of somethings being g is increased if that thing is f. in this way, salmon was able to respond to difficult noted earlier by scriven-we may explain the fact that the mayor has paresis by nothing that previously had untreated syphilis, even though the frequency with which untreated syphilitics contract paresis is small. on salmons account, nothing that the mayor had untreated syphilis gives an enormous boost to the probability of his having paresis, raising it from the baseline figure of close to 0 to about 15 percent.4 picking up the piecessalmons account of the explanation was deliberately motivated by the felt need to allow for explanations in the indeterministic contexts of contemporary physics. his approach dovetailed neatly with attempts, like those of patrick suppes, to fashion a conception of causality that would no long be restricted to deterniminstic situations. from 1970 to the present, one important strand in contemporary theories of explanation has taken explanation to consist in delineating the causation that will define causal relations in statistical terms. the simplest account of probabilistic causality would propose that a is causally relevant to b just in case p(a)not equal to p(b). unfortunately, this account is too simple. as hans reichenbach pointed out in the 1950s the inequality will obtain when a and b both effects of a common cause. thus further conditions must be imposed the statistical relations constitutive of probabilistic causation.since the 1970s a number of different proposals to inherit the position of orthodoxy once occupied by the covering-law model. most popular have been causal approaches to explanation, and, initially, proposals to ground explanation in a detailed conception of probabilistic causality promised to answer the four principal difficulties outlined above. however, it has proved remarkably difficult to work out a satisfactory account of explanation along these lines, and a number of critiques, most notably that by nancy cartwright, have cast doubt on the viability of the enterprise. faced with powerful objections, champions of causal approaches to explanation have pursued one of two options. one is to continue to honour humean concerns about the causal relation, and to seek an analysis of causation that will not make use of metaphysical notions that empiricists consider dubious. the most thorough attempt to carry out this programme has been undertaken by wesley salmon, who has attempted to develop reichenbachs account of causation in terms of the fundamental notion of mark transmission. the alternative approach is to declare victory by taking some causal notion as an unanalysed primitive, resisting humean scruples about how we might know how to apply this notion as misguided (perhaps the proposals of humphreys and cartwright should be viewed as embodying this approach).one evident attraction of the causal programme is that it provides an immediate response to the problem posed by the asymmetries of explanation. however, not all current theories view explanation as a matter of tracing causes. in recently years, bas van frassen, peter achinstein and peter railton have all made important contributions to the pragmatics of explanation, and the first two authors have defended the view that the enterprise of seeking substantive necessary conditions that apply across all contexts in which people seek and give explanations is misguided. the danger is that such pragmatic theories of explanation reduce the enterprise to triviality. for any explanation-seeking question and any proposition we choose, it seems that we can construct a context in which that proposition is licensed as an adequate explanatory answer to that question.a third cluster positions stays close to the covering-law models conception of explanations as arguments, proposing that explanatory argument are not distinguished singly but emerge from the best way of systematizing our body of knowledge. michael friedman and philip kitcher have developed accounts of explanation that take arguments to be explanatory if they belong to a system of arguments that best unifies our beliefs. one virtue of this approach is its ready unification and for addressing the asymmetries of explanation.the present debate echoes themes from earlier chapters in the history of philosophy. humes scruples about causation loom behind some efforts to articulate causal theories of explanation, those who oppose hume on causation sometimes seem to harken back to aris
温馨提示
- 1. 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
- 2. 本站的文档不包含任何第三方提供的附件图纸等,如果需要附件,请联系上传者。文件的所有权益归上传用户所有。
- 3. 本站RAR压缩包中若带图纸,网页内容里面会有图纸预览,若没有图纸预览就没有图纸。
- 4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
- 5. 人人文库网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对用户上传分享的文档内容本身不做任何修改或编辑,并不能对任何下载内容负责。
- 6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
- 7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。
最新文档
- 《2025年劳动合同解除协议范本》
- 2025年武汉劳动合同模板
- 2025年劳动合同制度与社会保障制度的融合与发展
- 搬运安全知识培训课件
- 精准选人用人新途径:村干部招聘面试题解读
- 工业互联网面试题库:各行业面试必 备
- 艺术学校面试经验分享:洛阳艺校面试题及应对策略
- 绿色能源领域求职者必 备:煤化工行业招聘面试题及答案解析
- 高级商务面试题库指南
- 高级生物信息学分析岗位面试题
- 昆明一中实验班数学试卷
- GB/T 18344-2025汽车维护、检测、诊断技术规范
- 2025年医院电子病历系统在医疗信息化中的应用优化与患者满意度报告
- 房屋应急维修管理办法
- 高考改革培训
- 中国电子艾灸仪行业投资分析及发展战略咨询报告
- 安全监理试题及试题答案
- 粮食机收减损培训课件
- 小学生编织手工课件
- 广西现代物流集团招聘笔试真题2024
- 2025餐饮劳动合同书 电子版
评论
0/150
提交评论