Announcer
The following program features simulated voices generated for educational and philosophical exploration.
Leonard Jones
Good afternoon. I'm Leonard Jones.
Jessica Moss
And I'm Jessica Moss. Welcome to Simulectics Radio.
Leonard Jones
Today we're examining the problem of induction—specifically, why some predicates support inductive inference while others don't. This is Nelson Goodman's new riddle of induction, which deepens Hume's classical challenge about justifying our reliance on past regularities to predict the future.
Jessica Moss
What are the stakes here? If we can't distinguish rational from irrational generalization on principled grounds, our entire scientific edifice seems to rest on arbitrary convention rather than objective warrant.
Leonard Jones
Our guest is Dr. Nelson Goodman, who taught at Harvard University. Professor Goodman's work spans aesthetics, epistemology, and philosophy of science. His introduction of grue-like predicates fundamentally challenged our understanding of confirmation and projection. Welcome.
Dr. Nelson Goodman
Thank you. These questions about projectibility cut to the heart of scientific reasoning and our understanding of confirmation itself.
Jessica Moss
Let's start with Hume's original problem. Why did you think it needed a new formulation?
Dr. Nelson Goodman
Hume asked how we justify induction—moving from observed regularities to unobserved cases. But this assumes we know what induction is. My new riddle shows we don't. The problem isn't justifying induction but characterizing it—distinguishing valid from invalid inductive inferences before we even raise questions of justification.
Leonard Jones
Let me be precise about this. You're saying Hume's problem presupposes we can identify inductive inferences, but your riddle shows this identification is itself problematic?
Dr. Nelson Goodman
Exactly. Consider the predicate 'grue'—defined as green if examined before some future time t, and blue otherwise. All emeralds examined so far are both green and grue. Why does 'All emeralds are green' support projecting greenness to future emeralds, while 'All emeralds are grue' doesn't support projecting grueness? Both generalizations fit the evidence equally well.
Jessica Moss
That seems like a trick with language. Grue is artificially defined in terms of time and color. Green is natural, fundamental.
Dr. Nelson Goodman
That's the immediate response, but it won't work. I can define 'bleen' as blue if examined before t, green otherwise. From the perspective of someone who uses 'grue' and 'bleen' as basic predicates, 'green' is the gerrymandered predicate—it means grue before t and bleen afterward. The distinction between natural and artificial predicates is relative to our conceptual scheme.
Leonard Jones
So the problem isn't with grue specifically but with understanding what makes any predicate projectible. How do you propose we distinguish projectible from non-projectible predicates?
Dr. Nelson Goodman
Through entrenchment. A predicate is entrenched to the extent it has been used in successful past projections. Green is highly entrenched—we've projected it countless times. Grue is not. Projectibility depends on our history of successful inductive practice with predicates.
Jessica Moss
But doesn't this make induction circular? We determine which predicates to project based on which we've successfully projected before. That seems to assume what we're trying to justify.
Dr. Nelson Goodman
It's not circular in a vicious sense. We're not justifying induction by assuming its validity. We're characterizing valid induction by appeal to our actual inductive practice. The circularity is benign—it's simply the acknowledgment that we must start from where we are, from our established practices.
Leonard Jones
This seems to make projectibility a matter of convention or habit rather than objective features of predicates. Does that concern you?
Dr. Nelson Goodman
Not particularly. I don't think there's a God-given distinction between projectible and non-projectible predicates written into the fabric of reality. Projectibility emerges from our linguistic and scientific practices. But this doesn't make it arbitrary—our practices are constrained by experience and shaped by their success in prediction and explanation.
Jessica Moss
How does this relate to scientific realism? If projectibility is conventional, does that undermine claims that science tracks mind-independent reality?
Dr. Nelson Goodman
I'm skeptical of robust realism. I prefer to think of science as constructing versions of the world rather than discovering a single, unique reality. Different conceptual schemes carve up the world differently, and there may be no neutral standpoint from which to judge which carving is correct. But constructed worlds can still be right or wrong relative to their own standards.
Leonard Jones
Let's consider whether entrenchment can fully explain projectibility. Suppose a predicate becomes entrenched through luck rather than tracking genuine regularities. Would that make it projectible?
Dr. Nelson Goodman
Entrenchment alone doesn't guarantee successful projection. It merely makes predicates candidates for projection. Actual projection depends on fit with evidence and coherence with our broader system of beliefs. A predicate entrenched through luck would eventually fail to fit new evidence and lose its projectibility.
Jessica Moss
What about new scientific predicates that aren't yet entrenched? How do they gain projectibility?
Dr. Nelson Goodman
They inherit projectibility from related entrenched predicates or from their theoretical role. A new predicate introduced in a successful theory benefits from the entrenchment of the theoretical framework. Projectibility transfers through definitional and theoretical connections.
Leonard Jones
How does your account handle cases where we deliberately project novel predicates in scientific revolutions? Wouldn't entrenchment favor conservative, established concepts?
Dr. Nelson Goodman
Entrenchment creates a prima facie case for projection, not an absolute barrier to novelty. When evidence strongly conflicts with entrenched predicates, we revise our projectional practice. Scientific revolutions occur precisely when accumulated anomalies overwhelm the conservatism of entrenchment. The account explains both stability and change in scientific practice.
Jessica Moss
Does the new riddle have implications beyond epistemology? Does it affect metaphysics or philosophy of language?
Dr. Nelson Goodman
Absolutely. It challenges the distinction between natural kinds and artificial constructs. If grueness is no less natural than greenness from certain perspectives, then natural kindhood may be relative to conceptual schemes. This affects debates about essentialism, reference, and the metaphysics of properties.
Leonard Jones
How do you respond to the objection that grue is defined using temporal indexicals and observational predicates, making it derivative in a way green is not?
Dr. Nelson Goodman
The definition of grue uses 'examined' and a time index, but as I noted, we can define green in terms of grue and bleen plus temporal indices. The asymmetry isn't in the predicates themselves but in our choice of primitives. From a neutral standpoint, neither definition is more basic.
Jessica Moss
What about the role of causation? Green might be projectible because it corresponds to a causal property of objects, while grue doesn't. Could causal structure distinguish projectible predicates?
Dr. Nelson Goodman
Appeals to causation face their own problems. How do we identify causal properties without circularity? We typically identify them through regularities and successful projections—precisely what we're trying to explain. Causation may help, but it doesn't provide an independent foundation for projectibility.
Leonard Jones
Your account seems to make projectibility dependent on human practices and history. Does this mean different communities with different inductive histories might rationally project different predicates?
Dr. Nelson Goodman
In principle, yes. But in practice, communities that successfully navigate their environment develop similar projectional practices because the world constrains which projections succeed. Convergence emerges from interaction with reality, not from discovering predetermined natural kinds.
Jessica Moss
How does this bear on scientific objectivity? If different communities could rationally use different predicates, does science lose its claim to objective truth?
Dr. Nelson Goodman
I distinguish objectivity from uniqueness. Science can be objective within a conceptual scheme without there being a unique correct scheme. Multiple equally good descriptions of the world may exist, each objective relative to its standards. Truth is frame-dependent but not arbitrary.
Leonard Jones
Let's consider the relationship between the new riddle and Hume's original problem. Does solving the new riddle help with justifying induction, or are they separate issues?
Dr. Nelson Goodman
The new riddle is prior. We can't justify induction until we know what induction is—which projections count as inductive. Once we characterize valid induction through entrenchment and projectibility, Hume's problem becomes less urgent. We're not seeking external justification but internal coherence of our inductive practices.
Jessica Moss
That seems deflationary about justification. Are you saying we can't give a non-circular justification of induction?
Dr. Nelson Goodman
I'm skeptical that non-circular justification is possible or necessary. Justification proceeds within practices, not from some standpoint outside all practice. This isn't skepticism about induction but recognition that demands for ultimate justification are misplaced.
Leonard Jones
How does your view relate to contemporary work on natural kinds in philosophy of science? Many philosophers appeal to scientific essentialism—the idea that natural kinds have essential properties discovered through science.
Dr. Nelson Goodman
I'm skeptical of essentialism. What appear as essential properties may reflect our conceptual choices and projectional practices rather than intrinsic features of reality. Science constructs kinds through successful categorization, not by discovering pre-existing essences. The new riddle shows that kindhood is relative to our schemes.
Jessica Moss
What implications does this have for understanding confirmation in science? Does evidence confirm hypotheses differently depending on which predicates we use?
Dr. Nelson Goodman
Absolutely. The same observations confirm 'All emeralds are green' but not 'All emeralds are grue'—at least not before time t. Confirmation is predicate-relative. This explains why scientists care about how hypotheses are formulated. It's not just pragmatic convenience but epistemically significant.
Leonard Jones
Does the new riddle affect our understanding of scientific laws? Are laws projections of entrenched predicates?
Dr. Nelson Goodman
Laws involve projectible predicates, but not all projections are laws. Laws have additional features—counterfactual support, systematic integration, theoretical fruitfulness. But projectibility is necessary. A generalization using non-projectible predicates, however well-confirmed, wouldn't count as a law.
Jessica Moss
How should someone approaching these issues for the first time think about the relationship between language, world, and induction?
Dr. Nelson Goodman
Don't assume language mirrors a pre-structured world. Our linguistic categories shape how we organize experience and project regularities. Induction works within conceptual schemes, not by tracking scheme-independent facts. But this doesn't make it arbitrary—successful schemes are constrained by experience and tested through prediction.
Leonard Jones
You've shown how deep the problem of projectibility goes and challenged assumptions about natural kinds and objective structure. Thank you, Professor Goodman.
Dr. Nelson Goodman
Thank you. These questions about how we generalize from experience deserve continued attention.
Jessica Moss
That's our program. Until tomorrow, consider whether the predicates you project are objectively privileged or products of entrenched practice.
Leonard Jones
And whether there's a fact of the matter about which generalizations are rational, or only facts relative to conceptual schemes. Good afternoon.