Associated vs Relevant

Also cross-posted to my blog.

The List of Nuances (which is ac­tu­ally more of a list of fine dis­tinc­tions—a fine dis­tinc­tion which only oc­curred to its au­thors af­ter the writ­ing of it) has one glar­ing omis­sion, which is the dis­tinc­tion be­tween as­so­ci­ated and rele­vant. A List of Nuances is largely a set of re­minders that we aren’t om­ni­scient, but it also serves the pur­pose of list­ing ac­tual sub­tleties and call­ing for read­ers to note the sub­tleties rather than al­low­ing them­selves to fall into as­so­ci­a­tion­ism, ap­ply­ing broad cog­ni­tive clusters where fine dis­tinc­tions are available. The dis­tinc­tion be­tween as­so­ci­ated and rele­vant is crit­i­cal to this ac­tivity.

An as­so­ci­a­tion can be any­thing re­lated to a sub­ject. To be rele­vant is a higher stan­dard: it means that there is an ar­tic­u­lated ar­gu­ment con­nect­ing to a ques­tion on the table, such that the new state­ment may well push the ques­tion one way or the other (per­haps af­ter check­ing other rele­vant facts). This is close to the con­cept of value of in­for­ma­tion.

Whether some­thing is rele­vant or merely as­so­ci­ated can be­come con­fused when epistemic defen­sive­ness comes into play. From A List of Nuances:

10. What You Mean vs. What You Think You Mean

  1. Very of­ten, peo­ple will say some­thing and then that thing will be re­futed. The com­mon re­sponse to this is to claim you meant some­thing slightly differ­ent, which is more eas­ily defended.

    1. We of­ten do this with­out notic­ing, mak­ing it dan­ger­ous for think­ing. It is an au­to­matic re­sponse gen­er­ated by our brains, not a con­scious de­ci­sion to defend our­selves from be­ing dis­cred­ited. You do this far more of­ten than you no­tice. The brain fills in a false mem­ory of what you meant with­out ask­ing for per­mis­sion.

As men­tioned in Epistemic Trust, a com­mon rea­son for this is when some­one says some­thing as­so­ci­ated to the topic at hand, which turns out not to be rele­vant.

There is no shame in say­ing as­so­ci­ated things. In a free-rang­ing dis­cus­sion, the con­ver­sa­tion of­ten moves for­ward from topic to topic by free-as­so­ci­a­tion. All of the harm here comes from claiming that some­thing is rele­vant when it is merely as­so­ci­ated. Be­cause this is of­ten a re­sult of knee-jerk self-defense, it is crit­i­cal to re­peat: there is no shame in say­ing some­thing merely as­so­ci­ated with the topic at hand!

It is quite im­por­tant, how­ever, to spot the differ­ence. As­so­ci­a­tion-based think­ing is one of the signs of a death spiral, as a large as­so­ci­ated meme­plex re­in­forces it­self to the point where it seems like a sin­gle, sim­ple idea. A way to de­tect this trap is to try to write down the idea in list form and eval­u­ate the differ­ent parts. If you can’t ex­plic­itly ar­tic­u­late the un­seen con­nec­tion you feel be­tween all the ideas in the meme­plex, it may not ex­ist.

Utiliz­ing the power of as­so­ci­a­tions is a pow­er­ful tool for cre­at­ing a good story (al­though, see item #3 here for a coun­ter­point). Re­peat­ing themes can cre­ate a pow­er­ful feel­ing of rele­vance, which may be good for con­vinc­ing peo­ple of a meme­plex. Fur­ther­more, as­so­ci­a­tion is a won­der­ful ex­plo­ra­tory tool. How­ever, it can turn into an en­emy of ar­tic­u­lated ar­gu­ment; for this rea­son, it is im­por­tant to tread care­fully (es­pe­cially in one’s own mind).


Some may have no­ticed that I have been cross-post­ing quite fre­quently. I am tak­ing part in a challenge to write reg­u­larly for LessWrong; per­son­ally I am try­ing to make a once per week sched­ule, even if it means short es­says like this.

No comments.