I think there’s a relevant difference here between being ignorant of actual data that you are aware exists (e.g. the color of hair), and being ignorant of the existence of alternative theories or models (e.g. possible alternative meanings of the word “color”).
That seemed to make sense to me at first, but I’m having a hard time actually finding a good dividing line to show the relevant difference, particularly since what seems like it can be model ignorance for one question can be data ignorance for another question.
For instance, here are possible statements about being ignorant about the question. “What is my spouse’s hair color?”
1: “I don’t know your spouse’s hair color.”
2: “I don’t know if your spouse has hair.”
In this context, 1 seems like data ignorance, and 2 would seem like model ignorance.
But given a different question “Does my spouse have hair?”
2 is data ignorance, and 1 doesn’t seem to be a well phrased response.
And there appear to be multiple levels of this as well: For instance, someone might not know whether or not I have a spouse.
What is the best way to handle this? Is it to simply try to keep track of the number of assumptions you are making at any given time? That seems like it might help, since in general, models are defined by certain assumptions.
I think there’s a relevant difference here between being ignorant of actual data that you are aware exists (e.g. the color of hair), and being ignorant of the existence of alternative theories or models (e.g. possible alternative meanings of the word “color”).
That seemed to make sense to me at first, but I’m having a hard time actually finding a good dividing line to show the relevant difference, particularly since what seems like it can be model ignorance for one question can be data ignorance for another question.
For instance, here are possible statements about being ignorant about the question. “What is my spouse’s hair color?”
1: “I don’t know your spouse’s hair color.”
2: “I don’t know if your spouse has hair.”
In this context, 1 seems like data ignorance, and 2 would seem like model ignorance.
But given a different question “Does my spouse have hair?”
2 is data ignorance, and 1 doesn’t seem to be a well phrased response.
And there appear to be multiple levels of this as well: For instance, someone might not know whether or not I have a spouse.
What is the best way to handle this? Is it to simply try to keep track of the number of assumptions you are making at any given time? That seems like it might help, since in general, models are defined by certain assumptions.