## MV Dialog

**The Shakespeare Conference: SHK 26.3781 Thursday, 27 August 2015**

[1] From: Gabriel Egan <This email address is being protected from spambots. You need JavaScript enabled to view it.>

Date: August 26, 2015 at 4:25:11 PM EDT

Subject: Re: SHAKSPER MV Dialog

[2] From: Pervez Rizvi <This email address is being protected from spambots. You need JavaScript enabled to view it.>

Date: August 27, 2015 at 4:50:17 AM EDT

Subject: Re: MV Dialog

[1]-----------------------------------------------------------------

From: Gabriel Egan <This email address is being protected from spambots. You need JavaScript enabled to view it.>

Date: August 26, 2015 at 4:25:11 PM EDT

Subject: Re: SHAKSPER MV Dialog

SHAKSPERians whose eyes are glazing over at Jim Carroll’s use of such terms as “The Chebyshev Inequality” should still be able to spot the logical incoherence of the following statements:

The probability that the number of occurrences of “man” is greater than say, 37, is the standard deviation squared divided by 37 squared, or 0.021, assuming the standard deviation is 5.4. So the fact that the actual value is 38 appears significant to me.

In the first sentence, the likelihood of the number of occurrences of “man” being greater than 37 is calculated as a small probability (0.021, around a 1-in-50 chance). In the second sentence the “actual value” of the number of occurrences of “man” is revealed to be 38.

If the “actual value” is 38 then the first sentence, as phrased, has no meaning. You cannot in any meaningful way calculate the likelihood of an outcome that has already come out. It’s like trying to calculate the likelihood that it rained here yesterday: the notion of likelihood simply doesn’t apply, since we already know whether it did.

We could ask this question: “Supposing that I didn’t know whether it rained here yesterday, what figure would I have given you the day before yesterday as the likelihood that it was going to rain?” This is NOT the same as asking “what is the likelihood that it rained yesterday?”, but people who use the mathematics of probability frequently treat these questions as equivalent. Doing so makes them commit fundamental errors regarding chance.

What Carroll REALLY means is something like “Suppose I didn’t know the ‘actual value’ because I haven’t gone counting the words yet. Further suppose, as the so-called ‘null hypothesis’, that the ‘actual value’ is determined by chance rather than by something else (say, authorial choice). If only chance were determining the ‘actual value’, what is the likelihood that when we go ahead and count the words to get the ‘actual value’ we will find an ‘actual value’ greater than, say, 37?”

If Carroll were to rephrase his claims using such precise language, and in particular to state just what his “null hypotheses” are, then it would be possible to start making sense of his claims.

Gabriel Egan

[2]-------------------------------------------------------------

From: Pervez Rizvi <This email address is being protected from spambots. You need JavaScript enabled to view it.>

Date: August 27, 2015 at 4:50:17 AM EDT

Subject: Re: MV Dialog

Jim writes:

The Chebyshev inequality is easy to calculate for the results in question.

The Chebyshev inequality is not a formula that you can use to calculate things with. It’s a law that probabilities obey.

The probability that the number of occurrences of “man” is greater than say, 37, is the standard deviation squared divided by 37 squared, or 0.021, assuming the standard deviation is 5.4.

I think Jim’s just making things up now. The calculation he has done is obvious nonsense, as you can see by replacing 37 with any value less than the standard deviation, when the probability will exceed 100%. It’s time to call it a day.