Phil12_S11_Causal_explanation(5-17-2011)

Phil12_S11_Causal_explanation(5-17-2011) - Causal...

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Causal explanation Phil 12: Logic and Decision Making Spring 2011 UC San Diego 5/17/2011 Tuesday, May 17, 2011 Review • Cause: something which brings about or increases the likelihood of an effect • Correlations point to and give evidence of causal relations, but do not themselves demonstrate causation • - Correlation is symmetrical Causation is directional Differentiate variables when investigating causal hypotheses: - Independent variable: possible cause, what is manipulated - Dependent variable: possible effect, depends on the independent variable Tuesday, May 17, 2011 Review • Strongest evidence for causation comes from experiments - Manipulate the independent variable and detect effect on the dependent variable - Speak of manipulated independent variable • When manipulation is not possible: - Measured independent variable • In either case, the dependent variable is measured Tuesday, May 17, 2011 Review: Testing causing claims • Often requires operationalizing the variable or developing measures in terms of which we can secure data • To conﬁrm a causal hypothesis - Identify a prediction that would not be true if the causal hypothesis were not true • • If the prediction is true, reason by modus tollens to the truth of the hypothesis To falsify a causal hypothesis - Identify a prediction that would be true were the hypothesis true • Tuesday, May 17, 2011 If the prediction is not true, reason by modus tollens to the falsity of the hypothesis Clicker question Opinion question: Is causation something you can perceive? A. Yes B. No C. Not sure Tuesday, May 17, 2011 Seeing Causes • A tradition in philosophy, whose roots lie with David Hume, maintains that we never see causation, only events preceding other events • But some visual displays such as those developed by Albert Michotte are hard to see in any other way Tuesday, May 17, 2011 Describe what you see Tuesday, May 17, 2011 What is a cause? • Cause: A variable, some of whose values bring about or increase the value of the effect variable • Two more speciﬁc notions of cause: - The cause is sufﬁcient to bring about the effect Tuesday, May 17, 2011 The cause is necessary to bring about the effect Recall the Logic of Necessary and Sufﬁcient Conditions • If the score is tied, then we will play another round. - The scores’s being tied is sufﬁcient for playing another round. - Tuesday, May 17, 2011 Playing another round is necessary when the score is tied. Sufﬁcient causes • Examples of factors sufﬁcient to bring about an effect - Dead battery is sufﬁcient for car not starting - Placing water in a normally operating freezer is sufﬁcient for it freezing - Ingesting (enough) hemlock is sufﬁcient for dying - Increased exercise without eating more is sufﬁcient for weight loss. Tuesday, May 17, 2011 Clicker question Which of the following describes the value of sufﬁcient causes? A. Eliminating a sufﬁcient cause will prevent the effect from occurring B. Eliminating a sufﬁcient cause will ensure that the effect occurs C. Supplying a sufﬁcient cause will ensure that the effect occurs D. Supplying a sufﬁcient cause will prevent the effect from occurring Tuesday, May 17, 2011 The value of sufﬁcient causes • A sufﬁcient cause gives us a recipe for producing an effect we want - If you want someone to be unable to drive your car, completely run down the battery - If you want to lose weight, exercise Tuesday, May 17, 2011 The difﬁculty with sufﬁcient causes • For many conditions in which you think you have found a sufﬁcient cause, an exception can be found - If you take an antidote with your hemlock (should one be found), you might escape death - If you put salt in the water, it may not freeze even when temperature is less than 32℉ Tuesday, May 17, 2011 The difﬁculty with sufﬁcient causes • Few factors we identify as causes are really sufﬁcient to bring about their effect - They sufﬁce only in the context of background conditions that are assumed to be in place • Turning the ignition switch will start the car only if it is hooked up to the rest of the ignition system, there is an engine in the car, there is gas, oxygen is available, . . . • Often need to specify a conjunction of factors to arrive at a sufﬁcient cause Tuesday, May 17, 2011 and it is very difﬁcult to note all of them Necessary causes • Something that is necessary to produce an effect: - Sex is necessary for producing babies - Early exposure to language is necessary for normal language development Tuesday, May 17, 2011 Oxygen is necessary for combustion Herpes zoster is a necessary cause of chickenpox For want of a... For want of a nail, the shoe was lost, For want of the shoe, the horse was lost, For want of the horse, the rider was lost, For want of the rider, the battle was lost, For want of the battle, the kingdom was And all for the want of a horseshoe nail! Tuesday, May 17, 2011 The value of necessary causes • Provide a way of preventing something - Avoiding sex does prevent babies (and AIDs) - Eliminating oxygen does stop ﬁres - Keeping Herpes zoster away from you prevents chickenpox Tuesday, May 17, 2011 The difﬁculty with necessary causes • For many supposed necessary causes, there are alternatives - Tuesday, May 17, 2011 Sex is not strictly necessary for producing babies—in vitro fertilization can replace it Clicker question Putting a jar over a candle is: A. A necessary cause for the candle burning B. A necessary cause for putting out the candle C. A sufﬁcient cause for putting out the candle D. Both a necessary and sufﬁcient cause for putting out the candle Tuesday, May 17, 2011 Recap Quick description Sufﬁcient Causes A sufﬁcient cause can bring about an effect, (e.g. ordinarily, pulling a trigger of a loaded gun sufﬁces to ﬁre a gun). A necessary cause is a precondition Necessary that must obtain for the effect to Causes manifest, (e.g. guns need a hammer to ﬁre.) Tuesday, May 17, 2011 Value of identifying Inadequacy as a full characterization Practical difﬁculties of ‘cause’ If we know sufﬁcient causes, we can be empowered to bring about desired effects, (e.g. if we know how to ﬁre the gun, we can.) Events which we may want to call causes aren’t strictly speaking sufﬁcient, since we can imagine contexts in which the causes fail to sufﬁce, (e.g. pulling a trigger ﬁres a gun, but not if the bullets are duds). If we know necessary causes, we can prevent effects, (e.g. remove the hammer and the gun will not ﬁre). Necessary conditions are also difﬁcult to Preconditions don’t do identify since the anything (just because a scientiﬁc enterprise can gun has all its parts, reveal, unexpectedly, doesn’t mean it will that what was once ﬁre.) supposed necessary is not (pin ﬁring) Sufﬁcient conditions are difﬁcult to identify, and, perhaps, impossible to perfectly specify, since such a speciﬁcation would require an exhaustive description of the relevant background conditions. Partial or contributory causes • Tad: “The problem with our schools is the teachers. There are too many incompetent teachers who either don’t know how to teach, or just don’t care about teaching any more.” • Ruby: “I think parents are a big part of the problem too. Lots of parents fail to read to their children, never help them with homework, and don’t make sure their kids get to bed on time.” • Tad: “So you’re saying bad teachers have nothing to do with it?” • Ruby: “No, I’m saying parents have a lot to do with it too.” Tuesday, May 17, 2011 Partial or contributory causes • A factor that increases the likelihood of the event occurring but may be neither necessary nor sufﬁcient for the effect Tuesday, May 17, 2011 Viewing TV violence and aggressive behavior Icy roads and car accidents Genetic factors and heart attacks Fast food diet and heart attacks Vigorous exercise and heart attacks Uncle John Exception • My uncle John smoked two packs of cigarettes a day for 75 years, and he never got lung cancer. See, smoking doesn’t cause lung cancer! - This would be an effective counterexample if the claim were that smoking is a sufﬁcient cause of lung cancer - But if the claim is that smoking is a contributory cause, one or even many counterexamples are not telling • Tuesday, May 17, 2011 Rather, what one must do is show that over a population there is no increase in lung cancer among those who smoke Why care about contributory causes? • Even though changes in contributory causes cannot either: - Totally prevent the effect Bring about the effect by themselves • they can signiﬁcantly increase or decrease the likelihood of the effect Tuesday, May 17, 2011 Eliminating fast foods from your diet can reduce the risk of heart attack Proximate/ultimate causes • What was the cause of the Cedar Fire in 2003? • A hunter’s ﬂare? Tuesday, May 17, 2011 Proximate/ultimate causes • Failure to remove the dry brush through forest thinning? • Dry conditions Tuesday, May 17, 2011 Proximate/ultimate causes • The growth of seedlings into trees? • The big bang? Tuesday, May 17, 2011 Proximate/ultimate causes • There is generally a history of events, each of which plays a role in bringing about the event of interest • Sometimes we are interested in events in close proximity to the effect we wish to explain—proximate cause • Sometimes we are interested in events further back in the chain of causation—ultimate causes • Proximate and ultimate causes are not competitors - They are relevant for different explanatory projects Tuesday, May 17, 2011 Identifying causes Tuesday, May 17, 2011 John Stuart Mill • Methods for selecting actual causes among possible causes (before the development of statistics!) - Start with variables assumed to include the possible causes - Use correlation to separate actual causes from possible causes • - Failure of a putative cause to correlate with the effect in the right way indicates lack of causation • Tuesday, May 17, 2011 Mill did not have modern statistics available, so he used simple, eye-ball correlations Or better, our inability to ﬁnd the cause! Mill’s methods 1. Method of agreement 2. Method of difference 3. Method of concomitant variation 4. Method of residues Tuesday, May 17, 2011 Method of agreement • “If two or more instances of the phenomenon under investigation have only one circumstance in common, the circumstance in which alone all the instances agree, is the cause (or effect) of the given phenomenon.” • Find cases in which the effect has occurred - Determine if there is only one thing that they all share Tuesday, May 17, 2011 If there is, that is (the likely) cause Method of agreement A C Tuesday, May 17, 2011 B D E H G C E K Method of agreement • Example: some cities have markedly lower rates of tooth decay - Is there anything these cities share in common? • If so, that is the likely cause of lower rates of tooth decay Tuesday, May 17, 2011 Table for Method of Agreement Dental Free Fluoride Education Dental in Water Program Clinics High salaries for dentists Low rates of tooth decay Dullsville Yes No Yes No Yes Bedroom Town No Yes Yes Yes Yes Golfville No No Yes No Yes Megacity Yes Yes Yes No Yes Tuesday, May 17, 2011 Table for Method of Agreement Dental Free Fluoride Education Dental in Water Program Clinics High salaries for dentists Low rates of tooth decay Dullsville Yes No Yes No Yes Bedroom Town No Yes Yes Yes Yes Golfville No No Yes No Yes Megacity Yes Yes Yes No Yes Tuesday, May 17, 2011 Method of Agreement at Work • Five patients all show amnesia after brain injury: - Patient 1: damage to the prefrontal cortex and the hippocampus - Patient 2: damage to the hippocampus, amygdala, and entorhinal cortex - Patient 3: damage to the thalamus and hippocampus - Patient 5: damage to the hippocampus and amygdala Tuesday, May 17, 2011 Patient 4: damage to the prefrontal cortex, hypothalamus, hippocampus, and amygdala Method of difference • “If an instance in which the phenomena under investigation occurs and an instance in which it does not occur, have every circumstance in common save one, that one occurring only in the former, the circumstance in which alone the two instances differ, is the effect, or the cause, or an indispensable part of the cause, of the phenomenon.” Tuesday, May 17, 2011 Method of difference - 2 • Find two things that differ in that one has the effect and the other doesn’t - If there is only one factor on which they differ, that is the likely cause A C Tuesday, May 17, 2011 B D A E D X E B Method of Difference - 3 • Example: four people apply for a loan, and only two get are approved for a loan - Is there anything that differs between the people who got the loan and the people who didn’t? • If so, that’s the likely cause of their not getting the loan Tuesday, May 17, 2011 Table of Method of Difference Loan College Earn over Own Declared Education \$80K Business Bankruptcy Approved Victor Yes Yes No Yes No Crystal Yes Yes No No Yes Tad Yes Yes No No Yes Chin Yes Yes No Yes No Tuesday, May 17, 2011 Table of Method of Difference Loan College Earn over Own Declared Education \$80K Business Bankruptcy Approved Victor Yes Yes No Yes No Crystal Yes Yes No No Yes Tad Yes Yes No No Yes Chin Yes Yes No Yes No Tuesday, May 17, 2011 Example of Yellow Fever • Once Walter Reed suspected mosquitoes as the transmitter of yellow fever He had one set of volunteers sleep on the soiled clothes and beds of yellow fever patients in a room screened so that no mosquitoes could get in. None of these people contracted the disease. - • • He had another group of volunteers stay completely away from sick patients, except he let mosquitoes that had been allowed to feast ﬁrst on people sick with the disease bite the patients. These volunteers did get sick. - Tuesday, May 17, 2011 Joint method of agreement and difference • The methods of agreement and difference can be used jointly: - Find something in common amongst all cases where the effect appears - Find matches for all these cases except that they lack the effect and the common ingredient • Example: Five factory workers are found to be inefﬁcient relative to others who are doing the same work. - The efﬁcient workers and the inefﬁcient workers were found to be similar in all relevant ways except one: the inefﬁcient workers were not part of a proﬁt sharing plan. - Conclusion: proﬁt sharing causes efﬁciency. Tuesday, May 17, 2011 Method of concomitant variation • “Whatever phenomenon varies in any manner whenever another phenomenon varies in some particular manner is either a cause or an effect of that phenomenon or is connected with it through some fact of causation.” Amount Amount Amount of of of Water Fertilizer Sunlight Crop Yield Plot A 2 51 8 Plot B 14 3 45 12 Plot C Tuesday, May 17, 2011 13 12 4 46 16 Method of concomitant variation • “Whatever phenomenon varies in any manner whenever another phenomenon varies in some particular manner is either a cause or an effect of that phenomenon or is connected with it through some fact of causation.” Amount Amount Amount of of of Water Fertilizer Sunlight Crop Yield Plot A 2 51 8 Plot B 14 3 45 12 Plot C Tuesday, May 17, 2011 13 12 4 46 16 Method of residues • “Subduct from any phenomenon such part as is known by previous inductions to be the effect of certain antecedents, and the residue of the phenomenon is the effect of the remaining antecedents.” C1 C2 Tuesday, May 17, 2011 E1 C1 E1 E2 Method of residues - 2 • • Distinguish three features of a plant's growth pattern: The development of large, healthy green leaves The development of strong stems and root structure The production of fruit and ﬂowers Applying a 10-10-10 fertilizer (nitrogen-phosphorus-potassium) results in all three. - What causes what? We know that: • • Nitrogen promotes the healthy growth of leaves Potassium encourages the development of stronger stems and roots In addition to these, our fertilized plants also produce fruit and ﬂowers more proliﬁcally than usual. • - Since we know what caused the improved growth of leaves, stems, and roots we infer that the "residue", the increase in the number of fruit and ﬂowers, was caused by the phosphorus • Tuesday, May 17, 2011 What is Mill saying? • Once you have identiﬁed the plausible candidate causes: Correlation (of a simple matching sort) can isolate the actual (or at least a good candidate) cause • Assumption behind Mill’s methods: one and only one factor is the cause, and it is one you have considered - Tuesday, May 17, 2011 But beware, this can sometimes backﬁre: • A man “drank rye and water on the ﬁrst night and became drunk. On the second night, he drank scotch and water and became drunk again. On the third night, he got drunk on bourbon and water. He therefore decided that the water was the cause of his getting drunk because it was the common element each time.” (Christiansen, 1994, p. 76) Mill’s Methods: Probative, but not Deﬁnitive • As the alcohol and water example shows, Mill’s methods do not always correctly identify the cause - The true cause might go unnoticed - The causal structure might be complex, involving interactions of multiple factors Something might correlate with the effect but not be the cause • Nonetheless, Mill’s methods are useful in clarifying our understanding of cause and how we test for it • The development of modern statistics came after Mill and provided a much more potent tool for identifying the factors Mill was seeking to identify Tuesday, May 17, 2011 ...
View Full Document

{[ snackBarMessage ]}

Ask a homework question - tutors are online