If un­der­stand­ing oth­ers from their own per­spec­tives is epis­tem­i­cal­ly valu­able, mere­ly pos­sess­ing a con­cept has an epis­temic cost, even ide­al­iz­ing away our con­tin­gent lim­i­ta­tions.

Ex­tant at­tempts at defin­ing a cat­e­go­ry of epis­temic emo­tions are in­ad­e­quate. We pro­pose an al­ter­na­tive, ac­cord­ing to which what makes an emo­tion type epis­temic is its fit­ting­ness con­di­tions.

I show how con­cep­tu­al lim­i­ta­tions could lead to epis­temic dilem­mas and how an epis­temic dilem­mist can ap­peal to un­avoid­able puz­zle­ment as a sign of dilem­mas in the same way that moral dilem­mists ap­peal to un­avoid­able guilt.

Ac­tu­al­ly 𝜑-ing doesn’t im­ply abil­i­ty to 𝜑 when the right kind of ran­dom­ness is in­volved. This kind of ran­dom­ness is in­volved in the sam­pling meth­ods that many cog­ni­tive sci­en­tists think we ac­tu­al­ly use in learn­ing. So, con­tra Fodor, we can learn con­cepts by hy­poth­e­sis test­ing with­out cir­cu­lar­i­ty.

We can take con­cepts to be in­con­sis­tent with­out go­ing in for in­fer­en­tial­ism. Bi­lat­er­al­ism is all that’s need­ed. But a puz­zle re­mains: what’s de­fec­tive about in­con­sis­tent con­cepts?

To make sense of the per­mis­si­bil­i­ty of off­set­ting, the stan­dard de­on­to­log­i­cal con­straint against do­ing harm should be re­placed with one against un­off­set harm in­creas­es.

S stu­pe­fies A when A ac­cepts S’s as­ser­tion with­out un­der­stand­ing it. I ar­gue stu­pe­fy­ing is an im­por­tant means both for good (co­op­er­a­tive, joint­ly ra­tio­nal in­quiry) and bad (ma­nip­u­la­tion) that cur­rent mod­els of con­ver­sa­tion do not ac­count for.

Con­tra Kit Fine, truth­mak­er se­man­tics should de­fine ex­act truth­mak­ing in terms of in­ex­act truth­mak­ing, rather than vice ver­sa.

Uni­ver­sal en­tail­ments of coun­ter­fac­tu­al don­key sen­tences aren’t as uni­ver­sal as as­sumed in the lit­er­a­ture. This makes de­riv­ing them from a con­tex­tu­al­ly pro­vid­ed sim­i­lar­i­ty or­der­ing more at­trac­tive than de­riv­ing them from a se­man­ti­cal­ly-en­cod­ed as­sign­ment sen­si­tiv­i­ty.

Coun­teri­den­ti­cals (“If I were you...”) and re­lat­ed more or­di­nary coun­ter­fac­tu­als show we should re­ject the Krip­ke-Ka­plan or­tho­doxy about in­dex­i­cals. Lewisian coun­ter­part the­o­ry looks like the way to go.