Your description of the context is sort of in the right direction but not quite correct. Let's take an obviously flawed hypothetical "rape" study question, "Did you have sex because of alcohol when you didn't mean to?". You can use a similar thought experiment to show the flaws with this question; both partners could easily answer "yes". What's going on should be more obvious in this blatant case. I think you should read Ted's mail as explaining this problem and questioning how many flaws the actual study shared with this example.
As for the specifics, your "unable to give informed consent" is ambiguous; the relevant interpretation here is drunk enough to possibly qualify as "raped" in the classification OF THE STUDY, not the view of actual law. I think his comment about law was more of a side explanation, not the core of his argument (he's not saying that "as such" the statistics are faulty).
Your second objection fails because it does not distinguish between the assumptions of the study and the law. The question is whether the STUDY clearly refused to count events as rape with anything less than "entirely unable to function" standard. You can't blame Ted if he uses the study's assumptions when criticizing it.
Your third objection has a pretty basic logic error. You assume there would always be a side who obviously "initiated" sex. Two people have sex while drunk. Neither intended it before getting drunk, both regret it afterwards, and their memories of what exactly happened are likely hazy. Are you sure the study always correctly identified which side counts as "initiating" the sex, and did not count that side as "raped"?
"You're trying to invent some distinction that means he only believes what he wrote in certain circumstances"
Intentionally misleading rhetoric. I've said his statement was too general and inaccurate, and it's unlikely he would mean all the absurd consequences a literal interpretation would have in other contexts. Suppose you try to explain some feature of Linux, and make a remark in your explanation that could be interpreted to apply more generally. Now saying that "Matthew Garrett believes that Microsoft Windows does absurd thing X" would be an inaccurate description of what you said, and likely false, even if that would follow from a literal interpretation of your comments.
BTW I tried look up the study to check what exactly the phrasing of the question was; it seems that even the authors of the original Koss study later agreed that it really was flawed and produced inflated statistics.
Re: My overall take on the discussion
As for the specifics, your "unable to give informed consent" is ambiguous; the relevant interpretation here is drunk enough to possibly qualify as "raped" in the classification OF THE STUDY, not the view of actual law. I think his comment about law was more of a side explanation, not the core of his argument (he's not saying that "as such" the statistics are faulty).
Your second objection fails because it does not distinguish between the assumptions of the study and the law. The question is whether the STUDY clearly refused to count events as rape with anything less than "entirely unable to function" standard. You can't blame Ted if he uses the study's assumptions when criticizing it.
Your third objection has a pretty basic logic error. You assume there would always be a side who obviously "initiated" sex. Two people have sex while drunk. Neither intended it before getting drunk, both regret it afterwards, and their memories of what exactly happened are likely hazy. Are you sure the study always correctly identified which side counts as "initiating" the sex, and did not count that side as "raped"?
"You're trying to invent some distinction that means he only believes what he wrote in certain circumstances"
Intentionally misleading rhetoric. I've said his statement was too general and inaccurate, and it's unlikely he would mean all the absurd consequences a literal interpretation would have in other contexts. Suppose you try to explain some feature of Linux, and make a remark in your explanation that could be interpreted to apply more generally. Now saying that "Matthew Garrett believes that Microsoft Windows does absurd thing X" would be an inaccurate description of what you said, and likely false, even if that would follow from a literal interpretation of your comments.
BTW I tried look up the study to check what exactly the phrasing of the question was; it seems that even the authors of the original Koss study later agreed that it really was flawed and produced inflated statistics.