Weakening an argument involving Causation

[monicapolissk]

Could someone read through these notes and let me know if my line of reasoning is correct?

1. To Weaken a Causal Argument, we address the assumptions. We don’t need to completely destroy the argument - we just need to make the assumption LESS LIKELY to be true.

Possibility 1: Causation in the Assumption

If the argument ASSUMES that A Causes B, a weakener would make this assumption less likely:

Inconsistent With
- No Cause, Effect (Makes it less likely assumption is true. Makes alternative cause more likely.)
- Cause, No Effect (Makes it less likely assumption is true. Makes alternative cause more likely)
*Recall that these DO NOT DESTROY THE ARGUMENT. Unless otherwise stated, a causal conclusion/assumption leaves open the possibility that it is not true in Every Case.

Directly Introduce Alternative
- Introduce Alternate Cause (Make causation assumption less likely)


Possibility 2: Causation in the Conclusion

These should be addressed a bit differently. We should not attack the claim of causation itself, but the argument. If the argument is something like “A is correlated with B, therefore A Causes B” we need to attack the assumption that “correlation implies causation”.

- Could it be the reverse? B causes A
- Could a 3rd thing cause both (and that is why we see a correlation)

[ohthatpatrick]

Hmmm, I applaud your attempt to systematize this, but I felt like the formality of this thinking might help you often but keep you from being flexible at other times.

I didn't really disagree with anything you said, although I'll admit (as a 10 year veteran of this test) that I've never once made a distinction between causation in the assumption and causation in the conclusion.

When it comes to Flaw questions, we do have to be aware that it is legitimate to say that correlations do not prove causality. That could be a valid answer (although it is more often a trap answer these days, on an argument that doesn't have a causal problem ... LSAT knows how famous this slogan/flaw is, so they dangle it as a trap answer that 'sounds right' routinely).

My framework for causal arguments is pretty simple. This is the same mindset I use whether the template is:

X happened --> Thus, Y is the reason.
X and Y are correlated --> Thus, Y causes X.

1. Are there OTHER WAYS to explain the evidence?
2. How PLAUSIBLE is the Author's Way?

When it comes to #1's, the most common ideas are
- reverse causality (if it was a correlation where we didn't know which came first)
- some third factor (that either causes both or accompanies one and causes the other)
in rare cases,
- normalcy (sometimes the OTHER WAY is that "this is simply a naturally occurring phenomenon that doesn't need a special explanation"

When it comes to #2, the most common ideas are
- Covariation (whether cause/effect rise and fall in tandem) ... that's where all the cause/effect, no cause / no effect (strengtheners) and cause/no effect, no cause/effect (weakeners) come from

- Anything else that affects Plausibility (proximity, awareness, relevance, trustworthy data)

There's obviously overlap between #1 and 2, since bringing up alternative possible explanations seems to decrease the plausibility that the AUTHOR's explanation is right. But you can generally categorize every correct answer to a causal argument as dealing with

1. presenting or ruling out a DIFFERENT WAY to understand/explain/interpret the evidence
or
2. delving into more detail about the AUTHOR'S WAY to increase/decrease plausibility

[monicapolissk]

Hmmm, I applaud your attempt to systematize this, but I felt like the formality of this thinking might help you often but keep you from being flexible at other times.

I didn't really disagree with anything you said, although I'll admit (as a 10 year veteran of this test) that I've never once made a distinction between causation in the assumption and causation in the conclusion.

When it comes to Flaw questions, we do have to be aware that it is legitimate to say that correlations do not prove causality. That could be a valid answer (although it is more often a trap answer these days, on an argument that doesn't have a causal problem ... LSAT knows how famous this slogan/flaw is, so they dangle it as a trap answer that 'sounds right' routinely).

My framework for causal arguments is pretty simple. This is the same mindset I use whether the template is:

X happened --> Thus, Y is the reason.
X and Y are correlated --> Thus, Y causes X.

1. Are there OTHER WAYS to explain the evidence?
2. How PLAUSIBLE is the Author's Way?

When it comes to #1's, the most common ideas are
- reverse causality (if it was a correlation where we didn't know which came first)
- some third factor (that either causes both or accompanies one and causes the other)
in rare cases,
- normalcy (sometimes the OTHER WAY is that "this is simply a naturally occurring phenomenon that doesn't need a special explanation"

When it comes to #2, the most common ideas are
- Covariation (whether cause/effect rise and fall in tandem) ... that's where all the cause/effect, no cause / no effect (strengtheners) and cause/no effect, no cause/effect (weakeners) come from

- Anything else that affects Plausibility (proximity, awareness, relevance, trustworthy data)

There's obviously overlap between #1 and 2, since bringing up alternative possible explanations seems to decrease the plausibility that the AUTHOR's explanation is right. But you can generally categorize every correct answer to a causal argument as dealing with

1. presenting or ruling out a DIFFERENT WAY to understand/explain/interpret the evidence
or
2. delving into more detail about the AUTHOR'S WAY to increase/decrease plausibility

Hi Patrick,

Thank you so much for the response!

The reason that I made the distinction between "Causation in the Assumption" & "Causation in the Conclusion" is because (at least it seems this way to me) you can't directly attack the causation when it is in the conclusion. So, if an argument says "A corr B. Therefore, A causes B.", I understood that to mean any answer like "cause without effect" and "effect without cause" was wrong since both are consistent with the correlation premise.

Am I correct in thinking this way?

[ohthatpatrick]

Yeah, that would be incorrect.

You can absolutely Strengthen/Weaken with covariation answers (cause/effect, ~cause/~effect, cause/~effect, ~cause/effect) when we go
X and Y are correlated, thus X caused Y.

With those types of answers, you really just need to make sure that it's got some good punching strength. If it's just saying "SOMETIMES / SOME PEOPLE are the cause but not the effect", that puny stuff can't weaken, unless the conclusion said "X is guaranteed to cause Y."

Overall, I'm confused what you're thinking in saying "you can't attack causation when it's in the conclusion".

That's EXACTLY what we should attack!

We always attack the conclusion, no matter what type of claim it is, assuming we're doing an Assumption Family question.