After reading the abstract, I was thinking to myself, 'to get a disagreement rate like that, how can you possibly present a subject's own argument to them and have them think it was someone else's? Wouldn't they remember?' and was expecting some very clever experimental manipulation or condition.
Nope:
> However, for one of the syllogisms (the manipulated syllogism), instead of being truthfully reminded of their previous answer, participants were told that they had given an answer different from the one they had given: either the valid answer (if they had answered invalidly) or the most common invalid answer (if they had answered validly). Their own previous answer, and the argument that justified it, were presented as if they were those given by another participant. The external features of the presentation were strictly identical to those of the other four syllogisms (see Fig. 1 for an example of both conditions).
This explains half the comments I see these days on HN. More concerned with defending their point and criticizing people who disagree, than a genuine attempt at finding something actually true.
Every so often, in places where I can be seen as the sum of my history (especially places with reputation scores), I create a new account with zero history. It stops me self-censoring on the basis of not wanting to damage my own reputation.
It's ridiculous; it's a line of text on a screen, and everyone here is a stranger, but I still feel myself caring. Can't stop caring, but I can start from fresh every so often.
Hold on. Turn this around. You just read an entire paper about how arguments are completely unreliable, to the point of causing you to contradict yourself, and your response is to argue semantics and declare, "Hey, arguments are great! Let's put more things in the 'argument' category, and spend less effort trying to figure out more reliable and accurate ways of reasoning! Yay arguments, and by entailment, yay self-contradiction!"
Nope:
> However, for one of the syllogisms (the manipulated syllogism), instead of being truthfully reminded of their previous answer, participants were told that they had given an answer different from the one they had given: either the valid answer (if they had answered invalidly) or the most common invalid answer (if they had answered validly). Their own previous answer, and the argument that justified it, were presented as if they were those given by another participant. The external features of the presentation were strictly identical to those of the other four syllogisms (see Fig. 1 for an example of both conditions).