Note- Gruebleen ( Green/Blue) first appeared in James Joyce's Finnegans Wake.
All Emeralds, examined, are Green, the unexamined ones, not yet dug up, are presumed Green too.
Now introduce a new property Grue, which is defined as; Grue things are examined and Green or unexamined and Blue.
Not only are all examined Emeralds green, they are also because already examined Grue. The unexamined emeralds cannot be both green and grue, since if they are grue and unexamined they are blue.
So why is it okay to say ALL Emeralds (including unexamined ones) are Green, but not all Emeralds are Grue ?
THE property grue looks suspicious and gerrymandered, and apparently less fundamental than blue and green. But if we had the concepts of grue and bleen ( blue if examined, green if not), first, we could define green and blue in terms of them. Green would be defined as grue if examined, bleen if not. Blue, bleen if examined, grue if not.
We can also imagine circumstances that grue would be useful. Suppose there were a sort of precious stone which was sensitive to examination, when exposed to light changed from blue to green. We may suppose exposure to light produces a chemical change, which if inhibited would leave the stones blue. Then all the stones would be grue, not just the examined ones.
We do not believe emeralds behave like that, however. We have good reason therefore to believe that if an unexamined emerald had been examined it would have been green and not blue, so we do not believe unexamined emeralds are grue.
It must be pointed out that we could never know the colour of unexamined emeralds, as soon as we go and look at them, they are examined.
The examined emeralds count as grue only because they have been examined, and our background knowledge tells us that being examined doesn't change them from blue to green. In short, we cannot generalize in a simple way from cases we have examined to all other cases without relevant background information. We can properly make the generalization only if that background information entitles us to regard the examined cases as a representative sample with respect to the property in question.
The imagined stones which lost their blue colour when exposed to light were not a representative sample with respect to green , since the sample included none of those yet exposed to light.
That they have been examined however, does not rule out the emeralds as green as opposed to grue, since we know that the normal examination of emeralds has no physical or chemical effect on them.
What if we start with grue and defined green as grue and examined and blue as grue and unexamined? Now the emeralds we have are grue and green. What of the unexamined ones? Well it is a feature of grue things that they are blue if unexamined. Our sample is unrepresentative for grue since it contains only examined gems.
If there is nothing in our background information to help us determine whether examined cases are typical, then we should refrain from making any generalizations until we have done more investigation, so that we acquire an idea of what factors might skew our example.
Paradox devised by Nelson Goodman.
All Emeralds, examined, are Green, the unexamined ones, not yet dug up, are presumed Green too.
Now introduce a new property Grue, which is defined as; Grue things are examined and Green or unexamined and Blue.
Not only are all examined Emeralds green, they are also because already examined Grue. The unexamined emeralds cannot be both green and grue, since if they are grue and unexamined they are blue.
So why is it okay to say ALL Emeralds (including unexamined ones) are Green, but not all Emeralds are Grue ?
THE property grue looks suspicious and gerrymandered, and apparently less fundamental than blue and green. But if we had the concepts of grue and bleen ( blue if examined, green if not), first, we could define green and blue in terms of them. Green would be defined as grue if examined, bleen if not. Blue, bleen if examined, grue if not.
We can also imagine circumstances that grue would be useful. Suppose there were a sort of precious stone which was sensitive to examination, when exposed to light changed from blue to green. We may suppose exposure to light produces a chemical change, which if inhibited would leave the stones blue. Then all the stones would be grue, not just the examined ones.
We do not believe emeralds behave like that, however. We have good reason therefore to believe that if an unexamined emerald had been examined it would have been green and not blue, so we do not believe unexamined emeralds are grue.
It must be pointed out that we could never know the colour of unexamined emeralds, as soon as we go and look at them, they are examined.
The examined emeralds count as grue only because they have been examined, and our background knowledge tells us that being examined doesn't change them from blue to green. In short, we cannot generalize in a simple way from cases we have examined to all other cases without relevant background information. We can properly make the generalization only if that background information entitles us to regard the examined cases as a representative sample with respect to the property in question.
The imagined stones which lost their blue colour when exposed to light were not a representative sample with respect to green , since the sample included none of those yet exposed to light.
That they have been examined however, does not rule out the emeralds as green as opposed to grue, since we know that the normal examination of emeralds has no physical or chemical effect on them.
What if we start with grue and defined green as grue and examined and blue as grue and unexamined? Now the emeralds we have are grue and green. What of the unexamined ones? Well it is a feature of grue things that they are blue if unexamined. Our sample is unrepresentative for grue since it contains only examined gems.
If there is nothing in our background information to help us determine whether examined cases are typical, then we should refrain from making any generalizations until we have done more investigation, so that we acquire an idea of what factors might skew our example.
Paradox devised by Nelson Goodman.
Last edited: