green numbers give full details.     |    back to list of philosophers     |     unexpand these ideas

Ideas of R Keefe / P Smith, by Text

[British, fl. 1997, both of Cambridge University.]

1997 Intro: Theories of Vagueness
1 p.2 If someone is borderline tall, no further information is likely to resolve the question
     Full Idea: If Tek is borderline tall, the unclarity does not seem to be epistemic, because no amount of further information about his exact height (or the heights of others) could help us decide whether he is tall.
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 1)
     A reaction: One should add also that information about social conventions or conventions about the usage of the word 'tall' will not help either. It seems fairly obvious that God would not know whether Tek is tall, so the epistemic view is certainly counterintuitive.
1 p.2 Vague predicates involve uncertain properties, uncertain objects, and paradoxes of gradual change
     Full Idea: Three interrelated features of vague predicates such as 'tall', 'red', 'heap', 'child' are that they have borderline cases (application is uncertain), they lack well-defined extensions (objects are uncertain), and they're susceptible to sorites paradoxes.
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 1)
     A reaction: The issue will partly depend on what you think an object is: choose from bundles of properties, total denial, essential substance, or featureless substance with properties. The fungal infection of vagueness could creep in at any point, even the words.
1 p.5 Many vague predicates are multi-dimensional; 'big' involves height and volume; heaps include arrangement
     Full Idea: Many vague predicates are multi-dimensional. 'Big' of people depends on both height and volume; 'nice' does not even have clear dimensions; whether something is a 'heap' depends both the number of grains and their arrangement.
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 1)
     A reaction: Anyone who was hoping for a nice tidy theory for this problem should abandon hope at this point. Huge numbers of philosophical problems can be simplified by asking 'what exactly do you mean here?' (e.g. tall or bulky?).
1 p.6 The simplest approach, that vagueness is just ignorance, retains classical logic and semantics
     Full Idea: The simplest approach to vagueness is to retain classical logic and semantics. Borderline cases are either true or false, but we don't know which, and, despite appearances, vague predicates have well-defined extensions. Vagueness is ignorance.
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 1)
     A reaction: It seems to me that you must have a rather unhealthy attachment to the logicians' view of the world to take this line. It is the passion of the stamp collector, to want everything in sets, with neatly labelled properties, and inference lines marked out.
1 p.7 A third truth-value at borderlines might be 'indeterminate', or a value somewhere between 0 and 1
     Full Idea: One approach to predications in borderline cases is to say that they have a third truth value - 'neutral', 'indeterminate' or 'indefinite', leading to a three-valued logic. Or a degree theory, such as fuzzy logic, with infinite values between 0 and 1.
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 1)
     A reaction: This looks more like a strategy for computer programmers than for metaphysicians, as it doesn't seem to solve the difficulty of things to which no one can quite assign any value at all. Sometimes you can't be sure if an entity is vague.
1 p.7 Supervaluationism keeps true-or-false where precision can be produced, but not otherwise
     Full Idea: The supervaluationist view of vagueness is that 'tall' comes out true or false on all the ways in which we can make 'tall' precise. There is a gap for borderline cases, but 'tall or not-tall' is still true wherever you draw a boundary.
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 1)
     A reaction: [Kit Fine is the spokesperson for this; it preserves classical logic, but not semantics] This doesn't seem to solve the problem of vagueness, but it does (sort of) save the principle of excluded middle.
1 p.15 If there is a precise borderline area, that is not a case of vagueness
     Full Idea: If a predicate G has a sharply-bounded set of cases falling in between the positive and negative, this shows that merely having borderline cases is not sufficient for vagueness.
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 1)
     A reaction: Thus you might have 'pass', 'fail' and 'take the test again'. But there seem to be two cases in the border area: will decide later, and decision seems impossible. And the sharp boundaries may be quite arbitrary.
2 p.19 The epistemic view of vagueness must explain why we don't know the predicate boundary
     Full Idea: A key question for the epistemic view of vagueness is: why are we ignorant of the facts about where the boundaries of vague predicates lie?
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 2)
     A reaction: Presumably there is a range of answers, from laziness, to inability to afford the instruments, to limitations on human perception. At the limit, with physical objects, how do we tell whether it is us or the object which is afflicted with vagueness?
3 p.23 Vague statements lack truth value if attempts to make them precise fail
     Full Idea: The supervaluationist view of vagueness proposes that a sentence is true iff it is true on all precisifications, false iff false on all precisifications, and neither true nor false otherwise.
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 3)
     A reaction: This seems to be just a footnote to the Russell/Unger view, that logic works if the proposition is precise, but otherwise it is either just the mess of ordinary life, or the predicate doesn't apply at all.
3 p.30 Some of the principles of classical logic still fail with supervaluationism
     Full Idea: Supervaluationist logic (now with a 'definite' operator D) fails to preserve certain classical principles about consequence and rules of inference. For example, reduction ad absurdum, contraposition, the deduction theorem and argument by cases.
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 3)
     A reaction: The aim of supervaluationism was to try to preserve some classical logic, especially the law of excluded middle, in the face of problems of vagueness. More drastic views, like treating vagueness as irrelevant to logic, or the epistemic view, do better.
3 p.32 The semantics of supervaluation (e.g. disjunction and quantification) is not classical
     Full Idea: The semantics of supervaluational views is not classical. A disjunction can be true without either of its disjuncts being true, and an existential quantification can be true without any of its substitution instances being true.
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 3)
     A reaction: There is a vaguely plausible story here (either red or orange, but not definitely one nor tother; there exists an x, but which x it is is undecidable), but I think I will vote for this all being very very wrong.
3 p.33 Supervaluation misunderstands vagueness, treating it as a failure to make things precise
     Full Idea: Why should we think vague language is explained away by how things would be if it were made precise? Supervaluationism misrepresents vague expressions, as vague only because we have not bothered to make them precise.
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 3)
     A reaction: The theory still leaves a gap where vagueness is ineradicable, so the charge doesn't seem quite fair. Logicians always yearn for precision, but common speech enjoys wallowing in a sea of easy-going vagueness, which works fine.
4 p.43 People can't be placed in a precise order according to how 'nice' they are
     Full Idea: There is no complete ordering of people by niceness, and two people could be both fairly nice, nice to intermediate degrees, while there is no fact of the matter about who is the nicer.
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 4)
     A reaction: This is a difficulty if you are trying to decide vague predicates by awarding them degrees of truth. Attempts to place a precise value on 'nice' seem to miss the point, even more than utilitarian attempts to score happiness.
4 p.46 If truth-values for vagueness range from 0 to 1, there must be someone who is 'completely tall'
     Full Idea: Many-valued theories still seem to have a sharp boundary between sentences taking truth-value 1 and those taking value less than 1. So there is a last man in our sorites series who counts as 'completely tall'.
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 4)
     A reaction: Lovely. Completely nice, totally red, perfectly childlike, an utter mountain, one hundred per cent amused. The enterprise seems to have the same implausibility found in Bayesian approaches to assessing evidence.
4 p.47 How do we decide if my coat is red to degree 0.322 or 0.321?
     Full Idea: What could determine which is the correct function, settling that my coat is red to degree 0.322 rather than 0.321?
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 4)
     A reaction: It is not just the uncertainty of placing the coat on the scale. The two ends of the scale have all the indeterminacy of being red rather than orange (or, indeed, pink). You are struggling to find a spot on the ruler, when the ruler is placed vaguely.
5 p.51 Objects such as a cloud or Mount Everest seem to have fuzzy boundaries in nature
     Full Idea: A common intuition is that a vague object has indeterminate or fuzzy spatio-temporal boundaries, such as a cloud. Mount Everest can only have arbitrary boundaries placed around it, so in nature it must have fuzzy boundaries.
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 5)
     A reaction: We would have to respond by questioning whether Everest counts precisely as an 'object'. At the microscopic or subatomic level it seems that virtually everything has fuzzy boundaries. Maybe boundaries don't really exist.
5 p.55 S5 collapses iterated modalities (◊□P→□P, and ◊◊P→◊P)
     Full Idea: S5 collapses iterated modalities (so ◊□P → □P, and ◊◊P → ◊P).
     From: R Keefe / P Smith (Intro: Theories of Vagueness [1997], 5)
     A reaction: It is obvious why this might be controversial, and there seems to be a general preference for S4. There may be confusions of epistemic and ontic (and even semantic?) possibilities within a single string of modalities.