Her: "If spectrum means X in this context it doesn't apply."

Youre going to need to provide a citation.

The only part I can find similar to this is where she talks about getting away from human perception. Something like emission spectra are exactly the type of objective, light-frequency based measurements shes talking about, and still dont match the definition she asserts.

The use of spectrum was always as an analogy by all involved.

Youre going to need to support that a lot more strongly.

The title of the video, and title of this thread, specifically say "Is Sex a spectrum?" and "Sex as Binary or Spectrum", and the starting point of the video is Bill Nye saying that sex is a spectrum. Now youre claiming that the word spectrum is only an

*analogy* of something that is opposed to the binary?

Bill Nye is

*clearly* not talking in analogy, he is using the standard, common definition of the word meaning a range of related elements with some overlapping characteristics. See definition 2

here. He is saying "sex has a number of related and overlapping characteristics which can be measured on a continuum". As the video is presented as a refutation or discussion of Bill's comment, if she is using spectrum as an analogy then she has failed to address Nye's comment, as well as used an exceedingly unclear analogy. What is she using the word spectrum as an analogy

*for*?

Not to mention the general use of "sex is a spectrum" is also not used as analogy, rather it is used in the same sense Nye uses it, so if shes only speaking in analogy shes just speaking to a strawman. In discussing spectrum only as analogy, she would be discussing something that nobody says.

A discrete set of criteria is different from a discrete set of values. BMI is a function of height and weight. There are incalculable numbers of various heights and weights if we go down to a small enough scale, but there are only two factors that then result in a single BMI. The conflation of discrete factors with discrete measurements of those factors is to miss her point.

I disagree. I believe

*she* misspoke in conflating discrete factors with discrete measurements. She does

*say* discrete factors, but then immediately says it "would rule out the possibility of sex being a true continuum". Discrete

*measurements* rule out the possibility of a continuum of classification, but have nothing to do with the number of factors under consideration. If she meant to say discrete factors, then her point is a non-sequitur. That would be akin to saying "Height has only one measurable factor, therefore height doesnt exist on a continuum".

As to the cultural means of appraising sex, this was not covered in this video. She has touched on it before, but given the now focused form for this video series, I imagine it will be its own topic.

I agree and that was partly my point. She presented a conclusion that was at least partly related to content she hasnt yet presented. It is, as yet, an unsupported conclusion.

Showing that the shape of the distribution of traits lends itself to two peaks, where most people cluster due to primary and secondary sexual characteristics correlating along two groups.

I disagree. The video did not make any argument regarding the actual shape of the distribution, or any argument regarding where "most" people cluster. It took

*as a given* that if we arranged people in a distribution it would look like the graph presented.

For the sake of argument or example, I was happy to go along with the idea that people would generally be distributed around two local maximums and look something

*very roughly* like the graph. But when we get to the idea of "most" or conclusions based on the shape of the graph, thats completely unsupported territory. We dont know how much of an overlap there is between the two groups. We dont know the mean clusters, the deviation from the norm, whether the two peaks are similar in size and deviation, or any statistically relevant information to make that assessment.

We dont have even

*have* a rough approximation of a distribution function, to determine how primary and secondary factors are weighted against each other. If we were to make a distribution somehow, it could just as easily come out something like this:

There are still two peaks, but its not as easy to simply classify the graph into two obvious categories or how many people in the middle count as one side or the other, or fit in neither classification.

Depending on the weighting function used to distill the primary and secondary characteristics down to a single data point, you could have a huge variety of statistical distributions. The only non-controversial method of weighting a distribution function in a way that would approximate the example graph is to base it off a weighting of factors that

*assumes* the binary classification and weights factors with more overlap as less important - which is circular reasoning.

The binary classification is not pragmatic from a statistical modelling perspective because creating a

*statistical model* is not pragmatic.