The correct definition of Nelson Goodmans famous predicate grue has been the subject of some controversy. Goodmans original definition is somewhat difficult to find online because most people do not quote Goodman verbatim, but instead offer their own paraphrase. Below are exact quotations from Goodmans book Fact, Fiction, and Forecast. Differences in different editions are noted (however, I have not managed to get a copy of the 3rd edition yet). I give, in addition to the definition of grue, the definition of other related words.
Chapter III, section 4: It is the predicate grue and it applies to all things examined before t just in case they are green but to other things just in case they are blue. Then at time t we have, for each evidence statement asserting that a given emerald is green, a parallel evidence statement asserting that that emerald is grue. And the statements that emerald a is grue, that emerald b is grue, and so on, will each confirm the general hypothesis that all emeralds are grue. Thus according to our definition, the prediction is that all emeralds subsequently examined will be green and the prediction that all will be grue are alike confirmed by evidence statements describing the same observations. But if an emerald subsequently examined is grue, it is blue and hence not green.
Consider...the predicate bleen that applies to emeralds examined before time t just in case they are blue and to other emeralds just in case they are green.
Let emerose apply just to emeralds examined before time t, and to roses examined later.
Chapter IV, section 4: Suppose that the predicate grund applies just to all things examined up to a certain time t that are green and to all things not so examined that are round. [Note: The word just was omitted in the 2nd edition, and Suppose that was changed to Suppose, however, in the 4th edition.] Footnote [omitted in 4th edition]: All emeralds are grund is not, it must be remembered, equivalent to All emeralds are green and round. All emeralds may be grund without all being green, without all being round, and even without any emerald being both green and round.
A thing is grare if either green and examined before t, or not so examined and square. [4th edition only]
Let the predicate emeruby apply to emeralds examined for color before t and to rubies not examined before t.
I would also like to take this opportunity to state my opinion about the oft-debated question about whether, in a world in which all emeralds are grue, emeralds change color. It seems clear to me from the above quotations that emeralds do not change color in such a world. Note that Goodman is careful always to include the word examined in his definitions; for example, something is grue if it is examined before t, and also green; he does not say that it is grue if it is green before t.
If you are not yet convinced, let me offer the following argument. If I were Goodman,
on the verge of inventing the concept of grue, I would want to construct a maximally
perverse example of induction going wrong, to ram home my point. Now what would be
maximally perverse? If all the emeralds that have ever been examined have been green, then
one would normally think that induction would predict that all the emeralds that have not
yet been examined will also be green. It would be maximally perverse for induction to
predict the exact opposite, namely that all the emeralds that have not yet been
examined are in fact not green (and are blue, say). This line of reasoning leads
naturally to the definition of grue as meaning that in a world where all emeralds
are grue, all the green emeralds get discovered first, before some specific time t,
and all the blue emeralds get discovered after that. In particular, no emerald changes its
color at any time.
Now, it is true that after one understands Goodmans basic idea, one can devise more complicated versions of it. In particular, given a single, specific emerald that I examine occasionally, I might note that every time I have examined that particular emerald, it has been green. Induction, one would think, would predict that on every subsequent examination, the emerald will also be green. The maximally perverse prediction would be that on every subsequent examination, the emerald will not be green (and will be blue, say). That is, in conventional language, the emerald would change color. While this scenario does still exemplify the basic insight behind the concept of grue, I very much doubt that this was what Goodman was thinking when he first came up with his new insight. For one thing, if this were really what he had in mind, it would have been more natural for him to propose the statement, This emerald is grue, rather than All emeralds are grue.
A second point to note is that if one examines the philosophical literature on induction, one sees that it is customary to posit a scientist collecting information about a species by examining specimens; for example, Hempels raven paradox from the 1940s imagines us examining ravens one at a time and noting that they are black, or examining non-black things and noting that they are not ravens. Hempel does not invite us to imagine examining the same non-black item repeatedly in order to confirm that it never turns into a raven; that much is taken for granted. Similarly, I take Goodman to be implicitly assuming that we are gathering information about emeralds by examining them one at a time, rather than by examining the same emerald repeatedly.
Finally, if we take the color-changing version of the paradox seriously, then we must be careful about what it means for an emerald to be a certain color. Implicit in our usual concept of an emerald being green is that the green coloration of the emerald persists. However, the persistence of the color is precisely what is being called into question by a color-changing version of the paradox. Therefore, to even state this version of the paradox properly, we should really start with the concept of being green at a particular time, and then be careful to define what we mean when we simply say that an emerald is green without explicitly mentioning time. That Goodman did not take pains to be precise about this point is, I believe, evidence that he did not have a color-changing version of the paradox in mind. He was taking for granted that the color of an emerald persists, so that one can speak of blue emeralds and green emeralds without further explanation.
In short, the color-changing version of the grue paradox is more complex and confusing and I am convinced that Goodman originally had in mind the simpler and crisper version in which individual emeralds do not change color.
Note added December 2017: I am grateful to Beppe Brivec, who provided the information below.
In the Foreword to chapter 8 induction of his book
Problems and Projects (page 359) Goodman wrote:
Occasionally grue is given some different interpretation. For example, in Positionality and Pictures and in the paper it discusses, grue is taken to apply not to entire enduring entities but rather to green time-slices examined before 2000 A.D. and to blue time-slices not so examined. Basically the same riddle arises and the shift in interpretation may often go unnoticed. Contrary to a common misunderstanding, however, the interpretation of grue originally given in Fact, Fiction, and Forecast (and in the above paragraph) does not require that a thing change from green to blue in order to remain grue, or from blue to green in order to remain bleen [Foreword to chapter 8 Induction of the Goodmans book Problems and Projects, page 359].