>>=
@noologist.bsky.social
I reached out just for the connection
but when she wove her hand in mine
my rushing heart beat out a code
now writ behind my eyes
an order cut into my dreams
to blacken out the skies
I post about computational cognitive science (she/her 🧡🤍💜)
but when she wove her hand in mine
my rushing heart beat out a code
now writ behind my eyes
an order cut into my dreams
to blacken out the skies
I post about computational cognitive science (she/her 🧡🤍💜)
The best pickup lines reference obscure early Oshii projects about losing his religion amid the failure of '68 and also possibly his sexual victimization (?) in slightly distressing style.
November 13, 2025 at 9:42 AM
The best pickup lines reference obscure early Oshii projects about losing his religion amid the failure of '68 and also possibly his sexual victimization (?) in slightly distressing style.
Melchior is unquestionably aspec until the Word of God(dess) says otherwise. His only lust is to OPPOSE.
November 13, 2025 at 7:57 AM
Melchior is unquestionably aspec until the Word of God(dess) says otherwise. His only lust is to OPPOSE.
(I am thinking of scientific inference as intuitionistic; i.e., without the law of the excluded middle for all sentences "not not p". This makes the expression of what we are doing in scientific inference a little more concise, although we could do it with LEM with a slightly different construction)
November 13, 2025 at 2:38 AM
(I am thinking of scientific inference as intuitionistic; i.e., without the law of the excluded middle for all sentences "not not p". This makes the expression of what we are doing in scientific inference a little more concise, although we could do it with LEM with a slightly different construction)
These sorts of subtler points are hard to understand even for many seasoned researchers, and especially for students, but are really important to get because of the sheer difficulty of making inferences of any kind in cognitive science.
November 12, 2025 at 9:06 PM
These sorts of subtler points are hard to understand even for many seasoned researchers, and especially for students, but are really important to get because of the sheer difficulty of making inferences of any kind in cognitive science.
Even if the hypothesis is not disconfirmed, there are always infinitely many other hypotheses that would be consistent with observation, and similarly if it is, there are infinitely many other sentences than the null which may be true. Neither is accepted by experiment.
November 12, 2025 at 9:06 PM
Even if the hypothesis is not disconfirmed, there are always infinitely many other hypotheses that would be consistent with observation, and similarly if it is, there are infinitely many other sentences than the null which may be true. Neither is accepted by experiment.
But we can, with some fallibility, say that observation is consistent or inconsistent with the hypothesis.
November 12, 2025 at 9:06 PM
But we can, with some fallibility, say that observation is consistent or inconsistent with the hypothesis.
Instead, a null entails that the hypothesis is false, a null rejection connotes that the hypothesis is not false, and a failure to reject connotes that the hypothesis is not not false. We cannot say exactly that the hypothesis is true or false.
November 12, 2025 at 9:06 PM
Instead, a null entails that the hypothesis is false, a null rejection connotes that the hypothesis is not false, and a failure to reject connotes that the hypothesis is not not false. We cannot say exactly that the hypothesis is true or false.
If the probabilities of error were 0, they would still not be the same. This is because observations of the world are not, without introduction of theory, amenable to induction (we have all read Hume after all). Causality may only be treated as yielding inductive set if we theorize it to do so.
November 12, 2025 at 9:06 PM
If the probabilities of error were 0, they would still not be the same. This is because observations of the world are not, without introduction of theory, amenable to induction (we have all read Hume after all). Causality may only be treated as yielding inductive set if we theorize it to do so.
Additional detail: the phrase "with p < x and beta < y we reject the null hypothesis" is the same as "with such and such a probability of types 1 and 2 error, observation is consistent with my hypothesis", but is not the same as saying "with some probability of error, my hypothesis is confirmed".
November 12, 2025 at 9:06 PM
Additional detail: the phrase "with p < x and beta < y we reject the null hypothesis" is the same as "with such and such a probability of types 1 and 2 error, observation is consistent with my hypothesis", but is not the same as saying "with some probability of error, my hypothesis is confirmed".
What is "derived" is more like an intuition that a certain inference is repeatable, even if the statistical practice used to get it in the exploratory phase was not kosher (as repeated analysis of the same data amplifies the likelihood that at least one apparent null rejection is a false positive).
November 12, 2025 at 9:06 PM
What is "derived" is more like an intuition that a certain inference is repeatable, even if the statistical practice used to get it in the exploratory phase was not kosher (as repeated analysis of the same data amplifies the likelihood that at least one apparent null rejection is a false positive).
I will add that I'm speaking in a strict and precise sense here. Empiricists often have the feeling that they do exploratory work and then derive hypotheses from the it, but the actual operation that's going on isn't "derivation" in the sense of deduction. It is inductive (or abductive).
November 12, 2025 at 9:06 PM
I will add that I'm speaking in a strict and precise sense here. Empiricists often have the feeling that they do exploratory work and then derive hypotheses from the it, but the actual operation that's going on isn't "derivation" in the sense of deduction. It is inductive (or abductive).
My new premises take the form of theoretical assertions that allow me to model how the population will change (or not change) between the moments when I am looking at it, so that I may predict additional observations than the ones I have made.
November 12, 2025 at 9:06 PM
My new premises take the form of theoretical assertions that allow me to model how the population will change (or not change) between the moments when I am looking at it, so that I may predict additional observations than the ones I have made.
Similarly, hypotheses cannot be derived from observation, in that if I see the population seems to be at this level today, it does not lead me to conclude that it will be at the some other level tomorrow, or even that it will be at the same level on the same day, unless I introduce new premises.
November 12, 2025 at 9:06 PM
Similarly, hypotheses cannot be derived from observation, in that if I see the population seems to be at this level today, it does not lead me to conclude that it will be at the some other level tomorrow, or even that it will be at the same level on the same day, unless I introduce new premises.
I am always measuring, not merely to report that I measured something, but to gain insight about the measured object, which I cannot measure exhaustively for all time. A measurement at a point in time says nothing in itself about other possible measurements at other points in time.
November 12, 2025 at 9:06 PM
I am always measuring, not merely to report that I measured something, but to gain insight about the measured object, which I cannot measure exhaustively for all time. A measurement at a point in time says nothing in itself about other possible measurements at other points in time.
To advocate the contrary is to go back to logical empiricism, which is a failed programme because mere conjunctions of observation and sentences that may be deduced from them are generally uninteresting.
November 12, 2025 at 9:06 PM
To advocate the contrary is to go back to logical empiricism, which is a failed programme because mere conjunctions of observation and sentences that may be deduced from them are generally uninteresting.
That model produces hypotheses; we need to validate it by designing it to predict other features of census data, and seeing if it's consistent with observation. All useful empirical insights come from generalizing from data in this way, which always requires saying more than can be per se deduced.
November 12, 2025 at 9:06 PM
That model produces hypotheses; we need to validate it by designing it to predict other features of census data, and seeing if it's consistent with observation. All useful empirical insights come from generalizing from data in this way, which always requires saying more than can be per se deduced.
All we can directly infer from the data is that person 1 responded at time t and so did person 2, and so did person 3. If we want to draw a conclusion about how many people live in the country, we must model how many people are expected to live in the country given recorded census responses.
November 12, 2025 at 9:06 PM
All we can directly infer from the data is that person 1 responded at time t and so did person 2, and so did person 3. If we want to draw a conclusion about how many people live in the country, we must model how many people are expected to live in the country given recorded census responses.
Census data is a good example. It's ordinary to have the intuition that a census just reports how many people live in the country, straight from the data, but this is wrong. The data as such record how many people were reported to have responded to the census at the time of collection.
November 12, 2025 at 9:06 PM
Census data is a good example. It's ordinary to have the intuition that a census just reports how many people live in the country, straight from the data, but this is wrong. The data as such record how many people were reported to have responded to the census at the time of collection.
It is the case per se that data are observations made at a particular point in time, and anything that can be directly inferred from them (even if we take them to be faultless) may be no stronger than the conjunction of all observations at a certain point in time.
November 12, 2025 at 9:06 PM
It is the case per se that data are observations made at a particular point in time, and anything that can be directly inferred from them (even if we take them to be faultless) may be no stronger than the conjunction of all observations at a certain point in time.
Not to say the American system is better lol just that Americans going all heart eyes over Canada is a little bit of a distortion. Shit's mostly only better here because the general mood of the country is more traditionally progressive and the Libs are more effective at holding power than the Dems
November 10, 2025 at 5:48 PM
Not to say the American system is better lol just that Americans going all heart eyes over Canada is a little bit of a distortion. Shit's mostly only better here because the general mood of the country is more traditionally progressive and the Libs are more effective at holding power than the Dems
Honestly the fact that we already did this bit 2 months ago and I forgot is making me question myself
If you’re using AI to do the creative part of your work for you I don’t just assume you are incompetent, I also believe you don’t have a soul.
October 29, 2025 at 5:40 AM
Honestly the fact that we already did this bit 2 months ago and I forgot is making me question myself