Book summary: Unlocking the Emotional Brain

If the the­sis in Un­lock­ing the Emo­tional Brain (UtEB) is even half-right, it may be one of the most im­por­tant books that I have read. Writ­ten by the psy­chother­a­pists Bruce Ecker, Robin Ti­cic and Lau­rel Hul­ley, it claims to offer a neu­ro­science-grounded, com­pre­hen­sive model of how effec­tive ther­apy works. In so do­ing, it also hap­pens to for­mu­late its the­ory in terms of be­lief up­dat­ing, helping ex­plain how the brain mod­els the world and what kinds of tech­niques al­low us to ac­tu­ally change our minds. Fur­ther­more, if UtEB is cor­rect, it also ex­plains why ra­tio­nal­ist tech­niques such as In­ter­nal Dou­ble Crux [1 2 3] work.

UtEB’s premise is that much if not most of our be­hav­ior is driven by emo­tional learn­ing. In­tense emo­tions gen­er­ate un­con­scious pre­dic­tive mod­els of how the world func­tions and what caused those emo­tions to oc­cur. The brain then uses those mod­els to guide our fu­ture be­hav­ior. Emo­tional is­sues and seem­ingly ir­ra­tional be­hav­iors are gen­er­ated from im­plicit world-mod­els (schemas) which have been formed in re­sponse to var­i­ous ex­ter­nal challenges. Each schema con­tains mem­o­ries re­lat­ing to times when the challenge has been en­coun­tered and men­tal struc­tures de­scribing both the prob­lem and a solu­tion to it.

Ac­cord­ing to the au­thors, the key for up­dat­ing such schemas in­volves a pro­cess of mem­ory re­con­soli­da­tion, origi­nally iden­ti­fied in neu­ro­science. The emo­tional brain’s learn­ings are usu­ally locked and not mod­ifi­able. How­ever, once an emo­tional schema is ac­ti­vated, it is pos­si­ble to si­mul­ta­neously bring into aware­ness knowl­edge con­tra­dict­ing the ac­tive schema. When this hap­pens, the in­for­ma­tion con­tained in the schema can be over­writ­ten by the new knowl­edge.

While I am not con­vinced that the au­thors are en­tirely right, many of the book’s claims definitely feel like they are point­ing in the right di­rec­tion. I will dis­cuss some of my caveats and reser­va­tions af­ter sum­ma­riz­ing some of the book’s claims in gen­eral. I also con­sider its model in the light of an is­sue of a psy­chol­ogy/​cog­ni­tive sci­ence jour­nal de­voted to dis­cussing a very similar hy­poth­e­sis.

Emo­tional learning

In UtEB’s model, emo­tional learn­ing forms the foun­da­tion of much of our be­hav­ior. It sets our ba­sic un­der­stand­ing about what situ­a­tions are safe or un­safe, de­sir­able or un­de­sir­able. The au­thors do not quite say it ex­plic­itly, but the gen­eral feel­ing I get is that the sub­cor­ti­cal emo­tional pro­cesses set many of the pri­ori­ties for what we want to achieve, with higher cog­ni­tive func­tions then try­ing to figure out how to achieve it—of­ten re­main­ing un­aware of what ex­actly they are do­ing.

UtEB’s first de­tailed ex­am­ple of an emo­tional schema comes from the case study of a man in his thir­ties they call Richard. He had been con­sis­tently suc­cess­ful and ad­mired at work, but still suffered from se­ri­ous self-doubt and low con­fi­dence at his job. On oc­ca­sions such as daily tech­ni­cal meet­ings, when he con­sid­ered say­ing some­thing, he ex­pe­rienced thoughts in­clud­ing “Who am I to think I know what’s right?”, “This could be wrong” and “Watch out—don’t go out on a limb”. Th­ese pre­vented him from ex­press­ing any opinions.

From the point of view of the au­thors, these thoughts have a definite cause—Richard has “emo­tional learn­ings ac­cord­ing to which it is adap­tively nec­es­sary to go into nega­tive thoughts and feel­ings to­wards [him­self].” The self-doubts are a strat­egy which his emo­tional brain has gen­er­ated for solv­ing some par­tic­u­lar prob­lem.

Richard’s ther­a­pist guided Richard to imag­ine what it would feel like if he was at one of his work meet­ings, made use­ful com­ments, and felt con­fi­dent in his knowl­edge while do­ing so. This was in­tended to elicit in­for­ma­tion about what Richard’s emo­tional brain pre­dicted would hap­pen if it failed to main­tain the strat­egy of self-doubt. The book in­cludes the fol­low­ing tran­script of what hap­pened af­ter Richard started imag­in­ing the scene as in­structed:

Richard: Now I’m feel­ing re­ally un­com­fortable, but-it’s in a differ­ent way.
Ther­a­pist: OK, let your­self feel it—this differ­ent dis­com­fort. [Pause.] See if any words come along with this un­com­fortable feel­ing.
Richard: [Pause.] Now they hate me.
Ther­a­pist: “Now they hate me.” Good. Keep go­ing: See if this re­ally un­com­fortable feel­ing can also tell you why they hate you now.
Richard: [Pause.] Hnh. Wow. It’s be­cause… now I’m… an ar­ro­gant ass­hole… like my father… a to­tally self-cen­tered, to­tally in­sen­si­tive know-it-all.
Ther­a­pist: Do you mean that hav­ing a feel­ing of con­fi­dence as you speak turns you into an ar­ro­gant ass­hole, like Dad?
Richard: Yeah, ex­actly. Wow.
Ther­a­pist: And how do you feel about be­ing like him in this way?
Richard: It’s hor­rible! It’s what I’ve always vowed not to be!

Richard had ex­pe­rienced his father as be­ing as­sertive as well as ob­nox­ious and hated. His emo­tional brain had iden­ti­fied this as a failure mode to be avoided: if you are as­sertive, then you are ob­nox­ious and will be hated. The solu­tion was to gen­er­ate feel­ings of doubt so as to stop him from be­ing too con­fi­dent. This caused him suffer­ing, but the pre­dic­tion of his emo­tional brain was that act­ing oth­er­wise would pro­duce even worse suffer­ing, as be­ing hated would be a ter­rible fate.

UtEB de­scribes Richard as hav­ing had the fol­low­ing kind of un­con­scious schema:

Per­cep­tual, emo­tional and so­matic mem­ory of origi­nal ex­pe­riences: his suffer­ing from his father’s heav­ily dom­i­nat­ing, hy­per-con­fi­dent self-ex­pres­sion, plus re­lated suffer­ing from un­met needs for fatherly ex­pres­sions of love, ac­cep­tance, un­der­stand­ing, val­i­da­tion. (This is the “raw data”; match­ing fea­tures in cur­rent situ­a­tions are trig­gers of the whole schema.)
A men­tal model or set of linked, learned con­structs op­er­at­ing as liv­ing knowl­edge of a prob­lem and a solu­tion:
The prob­lem: knowl­edge of a vuln­er­a­bil­ity to a spe­cific suffer­ing. Con­fi­dent as­sertive­ness in any de­gree in­flicts crush­ing op­pres­sion on oth­ers and is hated by them. I would be hor­rible like Dad and hated by oth­ers, as he is, if I as­serted my own knowl­edge or wishes con­fi­dently. (This is a model of how the world is, and cur­rent situ­a­tions that ap­pear rele­vant to this model are trig­gers of the whole schema.)
The solu­tion: knowl­edge of an ur­gent broad strat­egy and con­crete tac­tic(s) for avoid­ing that suffer­ing. Never ex­press any con­fi­dent as­sertive­ness, to avoid be­ing hor­rible and hated (gen­eral strat­egy and pro-symp­tom pur­pose), by vigilantly notic­ing any definite knowl­edge or opinions form­ing in my­self and block­ing them from ex­pres­sion by gen­er­at­ing po­tently self-doubt­ing, self-in­val­i­dat­ing thoughts (con­crete tac­tic and man­i­fested symp­tom).

Emo­tional schemas can be brought to light dur­ing a va­ri­ety of ways, in­clud­ing Fo­cus­ing, IFS, and imag­in­ing your­self do­ing some­thing and see­ing what you ex­pect to hap­pen as a re­sult.

But sup­pose that you do man­age to bring up a schema which seems wrong to you. What do you do then?

Me­mory re­con­soli­da­tion: up­dat­ing the emo­tional learning

The for­ma­tion of mem­ory traces in­volves con­soli­da­tion, when the mem­ory is first laid out in the brain; de­con­soli­da­tion, when an es­tab­lished mem­ory is “opened” and be­comes available for changes; and re­con­soli­da­tion, when a de­con­soli­dated mem­ory (along with pos­si­ble changes) is stored and be­comes frozen again. The term “re­con­soli­da­tion” is also used to re­fer to the gen­eral pro­cess from de­con­soli­da­tion to re­con­soli­da­tion; UtEB gen­er­ally ap­plies the term to mean the en­tire pro­cess. Un­less the con­text in­di­cates oth­er­wise, I do the same.

UtEB re­views some of the his­tory of mem­ory re­search. Un­til 1997, neu­ro­scien­tists be­lieved that past emo­tional learn­ing be­came per­ma­nently locked in the brain, so that mem­o­ries could only con­soli­date, never de- or re­con­soli­date. More re­cent re­search has in­di­cated that once a mem­ory be­comes ac­ti­vated, it is tem­porar­ily un­locked, al­low­ing it to be changed or erased.

Start­ing from 2004, new stud­ies sug­gested that ac­ti­va­tion alone is not suffi­cient to de­con­soli­date the mem­ory. The mem­o­ries are used to pre­dict that things will oc­cur in a similar fash­ion as they did pre­vi­ously. Be­sides just ac­ti­va­tion, there has to be a sig­nifi­cant mis­match be­tween what one ex­pe­riences and what the mem­ory sug­gests is about to hap­pen. The vi­o­la­tion of ex­pec­ta­tion can be qual­i­ta­tive (the pre­dicted out­come not oc­cur­ring at all) or quan­ti­ta­tive (the mag­ni­tude of the out­come not be­ing fully pre­dicted). In ei­ther case, it is this pre­dic­tion er­ror which trig­gers the de­con­soli­da­tion and sub­se­quent re­con­soli­da­tion.

The mem­ory era­sure seems to be spe­cific to the in­ter­pre­ta­tion from which the pre­dic­tion was pro­duced. For ex­am­ple, some­one who has had an ex­pe­rience of be­ing dis­liked may later ex­pe­rience be­ing liked. This may erase the emo­tional gen­er­al­iza­tion “I am in­her­ently dis­lik­able”, but it will not erase the mem­ory of the per­son also hav­ing been dis­liked.

Ap­plied re­con­soli­da­tion: an ex­am­ple of the schema up­date process

So, as­sum­ing that the model out­lined above is cor­rect, how does one ap­ply it in prac­tice?

From what we have dis­cussed so far, the es­sen­tial steps of eras­ing a learned be­lief (in­clud­ing an emo­tional schema) in­volves iden­ti­fy­ing it, ac­ti­vat­ing it, and then find­ing a mis­match be­tween its pre­dic­tion and re­al­ity.

The first difficulty is that the be­liefs in­volved with the schema are not nec­es­sar­ily con­sciously available at first. Richard knew that he suffered from a lack of self-es­teem, but he was not aware of its rea­son. The pro­cess started from him de­scribing in con­crete de­tails how this man­i­fested: as skep­ti­cal self-talk dur­ing a daily meet­ing.

As he was guided to imag­ine what would hap­pen if he didn’t have those thoughts and acted con­fi­dently, his ther­a­pist was seek­ing to re­trieve the im­plicit schema and bring it into con­scious­ness so that its con­tents would be­come available for ac­cess. Once it had been re­trieved, the ther­a­pist and Richard worked to­gether to ex­press the be­lief in the schema in max­i­mally emo­tional lan­guage:

“Feel­ing any con­fi­dence means I’m ar­ro­gant, self-cen­tered, and to­tally in­sen­si­tive like Dad, and peo­ple will hate me for it, so I’ve got to never feel con­fi­dent, ever.”

The au­thors have de­vel­oped a ther­a­peu­tic ap­proach called Co­her­ence Ther­apy, whose steps closely fol­low the steps of the mem­ory re­con­soli­da­tion pro­cess. The ex­am­ple of Richard is from this school of ther­apy.

In Co­her­ence Ther­apy (as well as re­lated ap­proaches, such as In­ter­nal Fam­ily Sys­tems), one ini­tially avoids any im­pulse to ar­gue or dis­prove with the re­trieved schema. This would risk it be­ing pushed away be­fore it has be­come suffi­ciently ac­ti­vated to al­low for re­con­soli­da­tion.

In­stead, one stays with it. Richard was given a card with the above phrase and in­structed to re­view it ev­ery day un­til the next ther­apy ses­sion, just feel­ing the ex­tent to which it felt true to him. This served to fur­ther in­te­grate ac­cess to the schema in ques­tion, mak­ing it bet­ter con­sciously available.

Two weeks later, Richard had fre­quently no­ticed his self-doubt, used it as a prompt for read­ing the card, and ex­pe­rienced its de­scrip­tion ring­ing true as a rea­son for his thoughts. When speak­ing with his ther­a­pist, he men­tioned a par­tic­u­lar event which had stuck in his mem­ory. In a re­cent meet­ing, he had thought of a solu­tion to a par­tic­u­lar prob­lem, but then kept quiet about it. A mo­ment later, an­other per­son had spo­ken up and sug­gested the same solu­tion in a con­fi­dent man­ner. Look­ing around, Richard had seen the per­son’s solu­tion and con­fi­dence be­ing re­ceived pos­i­tively by the oth­ers. Richard had been struck by how that re­ac­tion differed from what his schema pre­dicted would hap­pen if he had made the same sug­ges­tion in that tone.

Be­cause Richard had made the im­plicit as­sump­tions in his schema ex­plicit, he was able to con­sciously no­tice a situ­a­tion which seemed to vi­o­late those as­sump­tions: a pre­dic­tion mis­match. His ther­a­pist rec­og­nized this as a piece of con­tra­dic­tory knowl­edge which could be used to up­date the old schema. The ther­a­pist then guided Richard through a pro­cess in­tended to ac­ti­vate the old schema while bring­ing the con­tra­dic­tory in­for­ma­tion into aware­ness, trig­ger­ing a re­con­soli­da­tion pro­cess.

The ther­a­pist first in­structed Richard to men­tally bring him­self back to the situ­a­tion where he had just thought of the solu­tion, but held it back. To prop­erly ac­ti­vate the schema, the ther­a­pist guided Richard’s at­ten­tion to the pur­pose be­hind his re­luc­tance and Richard’s cer­tainty of any con­fi­dence mak­ing him dis­liked. Next, the ther­a­pist told Richard to re-live what hap­pened next: the other per­son mak­ing the same sug­ges­tion, and the other peo­ple in the room look­ing pleased rather than an­gry.

The book then has a tran­script of the ther­a­pist guid­ing Richard to re­peat this jux­ta­po­si­tion of the old schema and the dis­con­firm­ing ex­pe­rience (ital­i­cized brack­ets in the origi­nal):

Ther­a­pist: Stay with that. Stay with be­ing sur­prised at what you’re see­ing—sur­prised be­cause in your life, you’ve had such a definite know­ing that say­ing some­thing con­fi­dently to peo­ple will always come across like Dad, like an ob­nox­ious know-it-all, and peo­ple will hate that. That’s what you know, yet at the same time, here you’re see­ing that say­ing some­thing con­fi­dently isn’t always like Dad, and then peo­ple are fine with it. And it’s quite a sur­prise to know that. [That was an ex­plicit prompt­ing of an­other side-by-side ex­pe­rienc­ing of the two in­com­pat­i­ble know­ings, with the ther­a­pist ex­press­ing em­pa­thy for both, with no in­di­ca­tion of any fa­vor­ing of one know­ing over the other. The ther­a­pist paused for sev­eral sec­onds, then asked:] Does it feel true to de­scribe it like that? Your old know­ing right alongside this other new know­ing that’s so differ­ent?
Richard: [Quietly, seem­ing ab­sorbed in the ex­pe­rience.] Yeah.
Ther­a­pist: [Softly.] All along, it seemed to you that say­ing some­thing con­fi­dently could be done only in Dad’s dom­i­nat­ing way of do­ing it, and now sud­denly you’re see­ing that say­ing some­thing con­fi­dently can be done very differ­ently, and it feels fine to peo­ple. [This was an­other de­liber­ate rep­e­ti­tion of the same jux­ta­po­si­tion ex­pe­rience.]
Richard: Yeah.
Ther­a­pist: Mm-hm. [Silence for about 20 sec­onds.] So, how is it for you be in touch with both of these know­ings, the old one tel­ling you that any­thing said with con­fi­dence means be­ing like Dad, and the new one that knows you can be con­fi­dent in a way that feels okay to peo­ple? [Ask­ing this ques­tion re­peated the jux­ta­po­si­tion ex­pe­rience yet again, and, in ad­di­tion, the “how is it” por­tion of the ques­tion prompted Richard to view the ex­pe­rience with mind­ful or metacog­ni­tive aware­ness, while re­main­ing in the ex­pe­rience.]
Richard: It’s sort of weird. It’s like there’s this part of the world that I didn’t no­tice be­fore, even though it’s been right there.
Ther­a­pist: I’m in­trigued by how you put that. Sounds like a sig­nifi­cant shift for you.
Richard: Yeah, it is. Huh.
Ther­a­pist: You’re see­ing both now, the old part of the world and this other part of the world that’s new, even though it was right there all along. [That cued the jux­ta­po­si­tion ex­pe­rience for a fourth time, fol­lowed by silence for about 30 sec­onds.] So, keep see­ing both, the old part and the new part, when you open your eyes in a few sec­onds and come back into the room with me. [Richard soon opens his eyes and blinks a few times.] Can you keep see­ing both?
Richard: Yeah.
Ther­a­pist: What’s it like to see both and feel both now? [With the trans­for­ma­tion se­quence com­plete, this ques­tion be­gins the next step of ver­ifi­ca­tion— Step V—be­cause it probes for whether the tar­get learn­ing still ex­ists as an emo­tional ex­pe­rience.]
Richard: [Pause, then sud­den, glee­ful laugh­ter.] It’s kind of funny! Like, what? How could I think that? [This is an ini­tial marker in­di­cat­ing that the pro-symp­tom schema may have been suc­cess­fully dis­con­firmed, de­po­ten­ti­ated, and dis­solved by the trans­for­ma­tion se­quence.]
Ther­a­pist: Do you mean, how could you think that sim­ply say­ing what you know, or men­tion­ing some good idea that you’ve had, would make you seem ar­ro­gant, in­sen­si­tive and dom­i­nat­ing like Dad and be hated for it?
Richard: [Laugh­ing again.] Yeah!

After­wards, the ther­a­pist and Richard wrote a new card to­gether, which Richard was told to re­view daily:

All along it’s been so clear that if I con­fi­dently say what I know, I will always come across as ar­ro­gant, in­sen­si­tive, and dom­i­nat­ing like Dad, and be hated for it. And it’s so weird, look­ing around the room and see­ing that it doesn’t come across like that.

The pur­pose of the card was to provide ad­di­tional jux­ta­po­si­tion ex­pe­riences be­tween the old schema and the new knowl­edge. While the origi­nal trans­for­ma­tion se­quence might have been enough to elimi­nate the old schema, the schema might also have been stored in the con­text of many differ­ent situ­a­tions and con­tained in sev­eral mem­ory sys­tems. In such a situ­a­tion, fur­ther jux­ta­po­si­tions would have helped deal with it.

In a fol­low-up meet­ing, Richard re­ported hav­ing lost the feel­ings of self-doubt, and that speak­ing up no longer felt like it was any big deal. To ver­ify that the old schema re­ally had lost its power, the ther­a­pist tried de­liber­ately pro­vok­ing his old fears again:

Drop­ping his voice to a quieter tone, the ther­a­pist added, “But tell me, when you have some­thing to say and just say it, what about the dan­ger of com­ing across as a know-it-all, like Dad, and be­ing hated for that? What about your fear of that and how ur­gent it is to pro­tect your­self from that?” [...]
Richard took in the ques­tion, gazed at the ther­a­pist in silence for a few sec­onds, and then replied, “Well, I don’t know what to tell you. All I can say is, that doesn’t trou­ble me any more. And hear­ing you say it, it seems a lit­tle strange that it ever did—like, what was my prob­lem?”

Ap­plied re­con­soli­da­tion: the schema up­date pro­cess in general

Now that we have looked at a spe­cific ex­am­ple, we can look at a more gen­eral ver­sion of the pro­cess.

Ac­cess­ing sequence

In Co­her­ence Ther­apy, the ac­cess­ing se­quence is the pre­limi­nary phase of mak­ing both a per­son’s im­plicit schema and some dis­con­firm­ing knowl­edge ac­cessible, so that they can be used in the jux­ta­po­si­tion pro­cess:

  1. Symp­tom iden­ti­fi­ca­tion. Estab­lish­ing which spe­cific symp­toms the per­son re­gards as prob­le­matic, and when and where they man­i­fest. In Richard’s case, the gen­eral symp­tom was a lack of con­fi­dence, which speci­fi­cally man­i­fested as nega­tive self-talk in meet­ings.

  2. Retrieval of tar­get learn­ing. Bring­ing into ex­plicit aware­ness the pur­pose be­hind the symp­toms. This can then be used to guide the search for dis­con­firm­ing knowl­edge, as well as ac­cess­ing the origi­nal schema in or­der to re­con­soli­date it. In Richard’s case, the pur­pose was to avoid ex­press­ing con­fi­dence in a way that would make peo­ple hate him.

  3. Iden­ti­fi­ca­tion of dis­con­firm­ing knowl­edge. Iden­ti­fy­ing some past or pre­sent ex­pe­rience which di­rectly con­tra­dicts the origi­nal learn­ing. This knowl­edge does not nec­es­sar­ily need to feel “bet­ter” or “more pos­i­tive” than the old one, just as long as it is mu­tu­ally ex­clu­sive with the old one. In Richard’s case, the dis­con­firm­ing knowl­edge was the ex­pe­rience of his co-worker con­fi­dently propos­ing a solu­tion and be­ing well-re­ceived.

Era­sure sequence

Once both the tar­get schema and the dis­con­firm­ing knowl­edge are known, the era­sure steps can be ap­plied to up­date the learn­ing:

  1. Re­ac­ti­va­tion of the tar­get schema. Tap­ping into the felt truth of the origi­nal learn­ing, ex­pe­rienc­ing it as vividly as pos­si­ble.

  2. Ac­ti­va­tion of dis­con­firm­ing knowl­edge, mis­match­ing the tar­get schema. Ac­ti­vat­ing, at the same time, the con­tra­dic­tory be­lief and hav­ing the ex­pe­rience of si­mul­ta­neously be­liev­ing in two differ­ent things which can­not both be true.

  3. Rep­e­ti­tions of the tar­get-dis­con­fir­ma­tion pairing.

Some­thing that the au­thors em­pha­size is that when the tar­get schema is ac­ti­vated, there should be no at­tempt to ex­plic­itly ar­gue against it or dis­prove it, as this risks push­ing it down. Rather, the be­lief up­date hap­pens when one ex­pe­riences their old schema as vividly true, while also ex­pe­rienc­ing an en­tirely op­po­site be­lief as vividly true. It is the jux­ta­po­si­tion of be­liev­ing X and not-X at the same time, which trig­gers an in­built con­tra­dic­tion-de­tec­tion mechanism in the brain and forces a re­struc­tur­ing of one’s be­lief sys­tem to elimi­nate the in­con­sis­tency.

The book notes that this dis­t­in­guishes Co­her­ence Ther­apy from ap­proaches such as Cog­ni­tive Be­hav­ioral Ther­apy, which is premised on treat­ing some be­liefs as in­trin­si­cally ir­ra­tional and then seek­ing to dis­prove them. While UtEB does not go fur­ther into the com­par­i­son, I note that this is a com­mon com­plaint that I have heard of CBT: that by de­fault­ing to nega­tive emo­tions be­ing caused by be­lief dis­tor­tions, CBT risks be­lit­tling those nega­tive emo­tions which are ac­tu­ally pro­duced by cor­rect eval­u­a­tions of the world.

I would per­son­ally add that not only does treat­ing all of your be­liefs—in­clud­ing emo­tional ones—as pro­vi­sion­ally valid seem to be a re­quire­ment for ac­tu­ally up­dat­ing them, this ap­proach is also good ra­tio­nal­ity prac­tice. After all, you can only seek ev­i­dence to test a the­ory, not con­firm it.

If you no­tice differ­ent parts of your mind hav­ing con­flict­ing mod­els of how the world works, the cor­rect epistemic stance should be that you are try­ing to figure out which one is true—not priv­ileg­ing one of them as “more ra­tio­nal” and try­ing to dis­prove the other. Other­wise it will be un­avoid­able that your pre­con­cep­tion will cause you to dis­miss as false be­liefs which are ac­tu­ally true. (Of course, you can still rea­son­ably an­ti­ci­pate the be­lief up­date go­ing a par­tic­u­lar way—but you need to take se­ri­ously at least the pos­si­bil­ity that you will be shown wrong.)

Similar to what Eliezer pre­vi­ously sug­gested, this can ac­tu­ally be a re­lief. Not only would it be an er­ror as a mat­ter of prob­a­bil­ity the­ory to try to stack the deck to­wards re­ceiv­ing fa­vor­able ev­i­dence, do­ing so would also sab­o­tage the brain’s be­lief up­date pro­cess. So you might as well give up try­ing to do so, re­lax, and just let the ev­i­dence come in.

I spec­u­late that this limi­ta­tion might also be in place in part to help avoid the er­ror where you de­cide which one of two mod­els is more cor­rect, and then dis­card the other model en­tirely. Si­mul­ta­neously run­ning two con­tra­dic­tory schemas at the same time achieves good com­mu­ni­ca­tion within the brain, as it al­lows both of them to be prop­erly eval­u­ated and merged rather than one of them be­ing thrown away out­right. I sus­pect that in Richard’s case, the re­sult­ing pro­cess didn’t cause him to en­tirely dis­card the no­tion that some be­hav­iors will make him hated like his dad was—it just re­moved the over­gen­er­al­iza­tion which had been pro­duced by hav­ing too lit­tle train­ing data as the ba­sis of the schema.

Of course, this means that there does need to be some con­tra­dic­tory in­for­ma­tion available which could be used to dis­prove the origi­nal schema. One might have a schema for which no dis­con­fir­ma­tion is available be­cause it is cor­rect, or a schema which might or might not be cor­rect but which is mak­ing things worse and can­not eas­ily be dis­con­firmed. UtEB men­tions the ex­am­ple of a man, “Tó­mas”, who had a de­sire to be un­der­stood and val­i­dated by some­one im­por­tant in his life. Tó­mas re­marked that a pro­fes­sional ther­a­pist who was be­ing paid for his em­pa­thy could never fulfill that role. The up­date con­tra­dict­ing the schema that no­body in his life re­ally un­der­stood him, would have to come from some­one ac­tu­ally in his life.

Another is­sue that may pop up with the era­sure se­quence is that there is an­other schema which pre­dicts that, for what­ever rea­son, run­ning this trans­for­ma­tion may pro­duce ad­verse effects. In that case, one needs to ad­dress the ob­ject­ing schema first, es­sen­tially be car­ry­ing out the en­tire pro­cess on it be­fore re­turn­ing to the origi­nal steps. (This is similar to the phe­nomenon in e.g. In­ter­nal Fam­ily Sys­tems, where ob­ject­ing parts may show up and have their con­cerns ad­dressed be­fore work on the origi­nal part can pro­ceed.)

Ver­ifi­ca­tion step

Fi­nally, af­ter the era­sure se­quence has been run, one seeks to ver­ify that last­ing change has in­deed hap­pened and that the tar­get schema has been trans­formed. UtEB offers the fol­low­ing be­hav­ioral mark­ers as signs that a learn­ing which has pre­vi­ously gen­er­ated emo­tional re­sponses has in fact been erased:

  • “A spe­cific emo­tional re­ac­tion abruptly can no longer by re­ac­ti­vated by cues and trig­gers that formerly did so or by other stress­ful situ­a­tions.”

  • “Symp­toms of be­hav­ior, emo­tion, so­mat­ics, or thought that were ex­pres­sions of that emo­tional re­ac­tion also dis­ap­pear per­ma­nently.”

  • “Non-re­cur­rence of the emo­tional re­ac­tion and symp­toms con­tinues effortlessly and with­out coun­ter­ac­tive or pre­ven­tive mea­sures of any kind.”

The au­thors in­ter­pret cur­rent neu­ro­science to say that only mem­ory re­con­soli­da­tion can pro­duce these kinds of mark­ers. They can­not be pro­duced by coun­ter­ac­tive or com­pet­i­tive pro­cesses, such as try­ing to learn an op­po­site habit to re­place a neu­rotic be­hav­ior. Coun­ter­ac­tive pro­cesses are gen­er­ally frag­ile and sus­cep­ti­ble to re­lapse. When these mark­ers are ob­served in clini­cal work, UtEB ar­gues that one may in­fer that re­con­soli­da­tion has led to the origi­nal learn­ing be­ing re­placed.

Fur­ther examples

For ad­di­tional ex­am­ples of the schema up­date pro­cess, I recom­mend read­ing the book, which con­tains sev­eral more case stud­ies of is­sues which were dealt with us­ing this ap­proach. Here’s a brief sum­mary of the most de­tailed ones (note that some of these ex­am­ples are ac­tu­ally more de­tailed and in­clude ad­di­tional com­pli­ca­tions, such as more than one symp­tom-pro­duc­ing schema; I have only sum­ma­rized the most promi­nent ones to give a taste of them):

  • “Char­lotte”. Is­sue: ob­ses­sive at­tach­ment to a former lover. Schema: “It would be much bet­ter if I was merged with my lover”. Con­tra­dic­tory knowl­edge: The harm caused by not hav­ing bound­aries.

  • “Ted”. Is­sue: an in­abil­ity to hold a steady job and a gen­eral lack of suc­cess in life. Schema: “If my life is a mess, my father will be forced to ad­mit how badly he screwed up as a par­ent.” Con­tra­dic­tory knowl­edge: Real­iz­ing that Ted’s father would never ad­mit failure, no mat­ter what.

  • “Brenda”. Is­sue: stage fright when hav­ing a lead­ing role in an up­com­ing play. Schema: be­ing on the stage in front of an au­di­ence means be­ing un­able to get off, caus­ing hel­pless­ness similar to when Brenda was in a car with her al­co­holic father and couldn’t get off. Con­tra­dic­tory knowl­edge: re-imag­in­ing the scene and the way how Brenda could ac­tu­ally have got­ten out of the car.

  • “Travis”. Is­sue: in­abil­ity to ex­pe­rience in­ti­mate emo­tional close­ness in re­la­tion­ships. Schema: “No­body will pay at­ten­tion to how I feel or give me un­der­stand­ing for how I’m hurt­ing. I don’t mat­ter, and I’m all on my own.” Con­tra­dic­tory knowl­edge: the ther­a­pist’s em­pathic pres­ence and listen­ing.

  • “Regina”. Is­sue: strong anx­iety and panic dur­ing/​af­ter in­ter­act­ing with other peo­ple. Schema: “I’m ac­cept­able and lov­able only if I do ev­ery­thing perfectly.” Con­tra­dic­tory knowl­edge: Regina’s Un­cle Theo loves her re­gard­less of what im­perfec­tions she might have.

  • “Carol”. Is­sue: want­ing to avoid sex with her hus­band de­spite feel­ing emo­tion­ally close to him. Schema: en­gag­ing in any sex­u­al­ity means be­ing overtly sex­ual and harm­ing Carol’s daugh­ter, in the way that Carol was harmed by her mother’s overt sex­u­al­ity. Con­tra­dic­tory knowl­edge: once the schema was made con­scious, it ac­ti­vated the brain’s spon­ta­neous mis­match de­tec­tion mechanisms and started to feel silly.

As the last item sug­gests, some­times just mak­ing a schema ex­plicit is enough to start to dis­man­tle it. The au­thors sug­gest that the brain has a built-in de­tec­tion sys­tem which com­pares any con­sciously ex­pe­rienced be­liefs for in­con­sis­ten­cies with other things that a per­son knows, and can spon­ta­neously cre­ate jux­ta­po­si­tion ex­pe­riences by bring­ing up such in­con­sis­tent in­for­ma­tion. They sug­gest that ther­a­pies which are based on dig­ging up pre­vi­ously un­con­scious ma­te­rial, but which do not have an ex­plicit jux­ta­po­si­tion step, work to the ex­tent that the un­cov­ered ma­te­rial hap­pens to trig­ger this spon­ta­neous mis­match de­tec­tion. (We already saw this hap­pen­ing with Richard—once his un­der­ly­ing schema had been made con­scious, he was star­tled to later no­tice what seemed like a con­tra­dic­tion.)

One may note the con­nec­tion to the model in Con­scious­ness and the Brain that when some sub­sys­tem in the brain man­ages to ele­vate a men­tal ob­ject into the con­tent of con­scious­ness, mul­ti­ple sub­sys­tems will syn­chro­nize their pro­cess­ing around that ob­ject. If the ob­ject is an ex­plicit be­lief, then any sub­sys­tem which is pay­ing at­ten­tion to that ob­ject may pre­sum­ably de­tect in­con­sis­ten­cies with that sub­sys­tem’s own mod­els.

Be­sides these case stud­ies from Co­her­ence Ther­apy, the au­thors also an­a­lyze pub­lished case stud­ies from Ac­cel­er­ated Ex­pe­ri­en­tal Dy­namic Psy­chother­apy, Emo­tion-Fo­cused Ther­apy, Eye-Move­ment De­sen­si­ti­za­tion and Re­pro­cess­ing, and In­ter­per­sonal Neu­ro­biol­ogy. They try to show how these cases also car­ried out a jux­ta­po­si­tion pro­cess, even if the the­o­ret­i­cal frame­works of those ther­a­pies did not ex­plic­itly re­al­ize it. It is the claim of the au­thors that any ther­apy which causes last­ing emo­tional change does it through re­con­soli­da­tion. Fi­nally, the book con­tains four es­says from other ther­a­pists (us­ing Co­her­ence Ther­apy and EMDR), who an­a­lyze some of their own case stud­ies.

Eval­u­at­ing the book’s plausibility

Now that we have looked at the book’s claims, let’s look at whether we should be­lieve in them.

It is un­clear to me how re­li­able the neu­ro­science re­sults are; the au­thors cite a num­ber of stud­ies, but each in­di­vi­d­ual claim only refer­ences a rel­a­tively small num­ber of them.

On a brief look, I could not find any re­views or pa­pers that would have di­rectly made a crit­i­cal as­sess­ment of the book’s model. How­ever, I found some­thing that might be even bet­ter.

Be­hav­ioral and Brain Sciences is a re­spected jour­nal cov­er­ing sub­ject ar­eas across the cog­ni­tive sci­ences. BBS pub­lishes “tar­get ar­ti­cles” which pre­sent some kind of a the­sis or re­view about a par­tic­u­lar topic, to­gether with tens of brief com­men­taries which re­spond to the tar­get ar­ti­cle, and a fi­nal re­sponse by the tar­get ar­ti­cle’s au­thors to the com­men­taries.

In 2015, four pres­ti­gious (with a to­tal of 500 pub­lished re­search ar­ti­cles be­tween them) psy­chol­o­gists pub­lished a BBS tar­get ar­ti­cle, Me­mory re­con­soli­da­tion, emo­tional arousal, and the pro­cess of change in psy­chother­apy: New in­sights from brain sci­ence (Lane et al. 2015). While the ex­act model that they out­line has a num­ber of differ­ences from the UtEB model, the core idea is the same: that ther­a­peu­tic change from a wide va­ri­ety of ther­a­peu­tic ap­proaches, “in­clud­ing be­hav­ioral ther­apy, cog­ni­tive-be­hav­ioral ther­apy, emo­tion-fo­cused ther­apy, and psy­cho­dy­namic psy­chother­apy, re­sults from the up­dat­ing of prior emo­tional mem­o­ries through a pro­cess of re­con­soli­da­tion that in­cor­po­rates new emo­tional ex­pe­riences.”

One in­ter­est­ing differ­ence was that Lane et al. de­scribe emo­tional schemas some­what differ­ently. In their model, the schemas form mem­ory struc­tures with three mu­tu­ally in­te­grated com­po­nents: emo­tional re­sponses, epi­sodic/​au­to­bi­o­graph­i­cal mem­o­ries, and se­man­tic struc­tures (e.g. ab­stract be­liefs which gen­er­al­ize over the var­i­ous in­ci­dents, such as the claim that “peo­ple are un­trust­wor­thy”). Any of these com­po­nents can be used as an en­try point to the mem­ory struc­ture, and can po­ten­tially up­date the other com­po­nents through re­con­soli­da­tion. They hy­poth­e­size that differ­ent forms of ther­apy work by ac­cess­ing differ­ent types of com­po­nents: e.g. be­hav­ior ther­apy and emo­tion-fo­cused ther­apy ac­cess emo­tional re­sponses, con­ven­tional psy­cho­anal­y­sis uses ac­cess to bi­o­graph­i­cal mem­o­ries, and cog­ni­tive be­hav­ioral ther­apy ac­cesses se­man­tic struc­tures.

I read the tar­get ar­ti­cle, all the com­men­taries, and the re­sponses. Given the similar­i­ties be­tween Lane et al.’s model and the UtEB model, I think we can con­sider the re­sponses to Lane et al. to gen­er­ally offer a use­ful eval­u­a­tion of the UtEB model as well.

One sig­nifi­cant differ­ence which needs to be noted is that Lane et al.’s model of mem­ory re­con­soli­da­tion does not men­tion the re­quire­ment for a pre­dic­tion mis­match be­fore re­con­soli­da­tion can hap­pen. This was re­marked on in the re­sponse from UtEB’s au­thors. In their counter-re­sponse, Lane et al. noted UtEB’s model to be highly com­pat­i­ble with theirs, and re­marked that fur­ther re­search is needed to nail down the con­di­tions which make re­con­soli­da­tion the most effec­tive.

The other re­sponses to Lane et al. were mostly from psy­chol­o­gists, psy­chi­a­trists, and neu­ro­scien­tists, but also in­cluded the oc­ca­sional economist, philoso­pher, philol­o­gist and folk­lorist. Sev­eral of the re­sponses were gen­er­ally pos­i­tive and mostly wanted to con­tribute ad­di­tional de­tails or point out fu­ture re­search di­rec­tions.

How­ever, there were also a num­ber of skep­ti­cal re­sponses. A com­mon theme which emerged from sev­eral con­cerned the limi­ta­tions of the cur­rent neu­ro­science re­search on mem­ory re­con­soli­da­tion. In par­tic­u­lar, most of the stud­ies so far have been car­ried out on rats, and speci­fi­cally test­ing the elimi­na­tion of a fear re­sponse to elec­tric shocks. As one of the re­sponses points out, “nei­ther the stim­uli nor the sub­jects are gen­er­al­iz­able to the kind of rich au­to­bi­o­graph­i­cal mem­o­ries in­volved in ther­apy.” A num­ber also raised the ques­tion whether all ther­a­peu­tic change re­ally in­volves re­con­soli­da­tion, as op­posed to some re­lated mechanism, such as cre­at­ing new mem­ory struc­tures which com­pete with the origi­nal as op­posed to re­plac­ing it.

My non-ex­pert read­ing is that the crit­i­cal re­sponses are right in that a gap re­mains be­tween the clini­cal and be­hav­ioral find­ings on the other hand, and the neu­ro­science find­ings on the other. There are var­i­ous pat­terns which can be de­rived from psy­cholog­i­cal re­search and clini­cal ther­apy ex­pe­rience, and a small num­ber of neu­ro­science find­ings which es­tab­lish the ex­is­tence of some­thing that could ex­plain those pat­terns. How­ever, the neu­ro­science find­ings have only been es­tab­lished in a rather nar­row and limited con­text; the con­nec­tion be­tween them and the higher-level pat­terns is a plau­si­ble link, but it re­mains spec­u­la­tive nonethe­less.

Per­son­ally I con­sider the book’s model ten­ta­tively promis­ing, be­cause it seems to ex­plain many ob­ser­va­tions which I had in­de­pen­dently ar­rived at be­fore read­ing it. For ex­am­ple, I had no­ticed an in­ter­est­ing thing with anx­ieties, where I let e.g. a sen­sa­tion of so­cial anx­iety stay ac­tive in my mind, nei­ther ac­cept­ing it as truth nor push­ing it away while I went to do so­cial things. This would then cause the anx­iety to up­date, mak­ing me feel less anx­ious if it was in­deed the case that the so­cial in­ter­ac­tion was harm­less. This fits nicely to­gether with the frame­work of an ac­ti­vated mem­ory struc­ture be­com­ing open to re­con­soli­da­tion and then be­ing up­dated by a pre­dic­tion mis­match (the situ­a­tion not be­ing as bad as ex­pected).

Like­wise, in my post In­te­grat­ing dis­agree­ing sub­agents, I re­viewed a va­ri­ety of ra­tio­nal­ity and ther­a­peu­tic tech­niques, and sug­gested that they mostly worked ei­ther by merg­ing or com­bin­ing two ex­ist­ing mod­els that a per­son’s brain already had, or aug­ment­ing the ex­ist­ing mod­els by col­lect­ing ad­di­tional in­for­ma­tion.

In par­tic­u­lar, I con­sid­ered an ex­am­ple from cog­ni­tive be­hav­ioral ther­apy, where a man named Walter was feel­ing like he was im­pos­si­ble to be in a re­la­tion­ship with af­ter he had bro­ken up with his boyfriend. At the same time, he did not think that some­one else break­ing up with their part­ner was an in­di­ca­tion of them be­ing im­pos­si­ble to be in a re­la­tion­ship with. He and his ther­a­pist role-played an in­ter­ac­tion where the ther­a­pist pre­tended to be a friend who had re­cently bro­ken up, and Walter ex­plained why this did not make the friend a re­la­tion­ship failure. In the pro­cess of do­ing so, Walter sud­denly re­al­ized that he wasn’t a failure, ei­ther.

I com­mented:

Walter was asked whether he’d say some­thing harsh to a friend, and he said no, but that alone wasn’t enough to im­prove his con­di­tion. What did help was putting him in a po­si­tion where he had to re­ally think through the ar­gu­ments for why this is ir­ra­tional in or­der to con­vince his friend, and then, af­ter hav­ing for­mu­lated the ar­gu­ments once him­self, get con­vinced by them him­self.
In terms of our frame­work, we might say that a part of Walter’s mind con­tained a model which out­put a harsh judg­ment of him­self, while an­other part con­tained a model which would out­put a much less harsher judg­ment of some­one else who was in oth­er­wise iden­ti­cal cir­cum­stances. Just bring­ing up the ex­is­tence of this con­tra­dic­tion wasn’t enough to change it: it caused the con­tra­dic­tion to be no­ticed, but didn’t ac­ti­vate the rele­vant mod­els ex­ten­sively enough for their con­tents to be re­pro­cessed.
But when Walter had to role-play a situ­a­tion where he thought of him­self as ac­tu­ally talk­ing with a de­pressed friend, that re­quired him to more fully ac­ti­vate the non-judg­men­tal model and ap­ply it to the rele­vant situ­a­tion. This caused him to blend with the model, tak­ing its per­spec­tive as the truth. When that per­spec­tive was then prop­a­gated to the self-crit­i­cal model, the eas­iest way for the mind to re­solve the con­flict was sim­ply to al­ter the model pro­duc­ing the self-crit­i­cal thoughts.

This seems like a straight­for­ward in­stance of be­lief jux­ta­po­si­tion, and one where I ended up in­de­pen­dently de­riv­ing some­thing like UtEB’s mem­ory re­con­soli­da­tion model: I too noted that the rele­vant be­lief struc­tures need to be si­mul­ta­neously ac­ti­vated in the right way to al­low for the brain to re­vise one of them af­ter notic­ing the con­tra­dic­tion. In gen­eral, UtEB’s model of how things work rings true in my ex­pe­rience, mak­ing me in­clined to be­lieve that its de­scrip­tion of how ther­apy works is cor­rect, and that its model of how it is con­nected to neu­ro­science might also be.

UtEB and the sub­agent model

As many read­ers know, I have been writ­ing a se­quence of posts on multi-agent mod­els of mind. In Build­ing up to an In­ter­nal Fam­ily Sys­tems model, I sug­gested that the hu­man mind might con­tain some­thing like sub­agents which try to en­sure that past catas­tro­phes do not re­peat. In sub­agents, co­her­ence, and akra­sia in hu­mans, I sug­gested that be­hav­iors such as pro­cras­ti­na­tion, in­de­ci­sion, and seem­ingly in­con­sis­tent be­hav­ior re­sult from differ­ent sub­agents hav­ing dis­agree­ments over what to do.

As I already men­tioned, my post on in­te­grat­ing dis­agree­ing sub­agents took the model in the di­rec­tion of in­ter­pret­ing dis­agree­ing sub­agents as con­flict­ing be­liefs or mod­els within a per­son’s brain. Subagents, trauma and ra­tio­nal­ity fur­ther sug­gested that the ap­pear­ance of dras­ti­cally differ­ent per­son­al­ities within a sin­gle per­son might re­sult from un­in­te­grated mem­ory net­works, which re­sist in­te­gra­tion due to var­i­ous trau­matic ex­pe­riences.

This post has dis­cussed UtEB’s model of con­flict­ing emo­tional schemas in a way which fur­ther equates “sub­agents” with be­liefs—in this case, the var­i­ous schemas seem closely re­lated to what e.g. In­ter­nal Fam­ily Sys­tems calls “parts”. In many situ­a­tions, it is prob­a­bly fair to say that this is what sub­agents are.

That said, I think that while this cov­ers a very im­por­tant sub­set of sub­agents, not ev­ery­thing which I have been refer­ring to as a sub­agent falls straight­for­wardly un­der the be­lief-schema model. In sub­agents and neu­ral Tur­ing ma­chines as well as Against “Sys­tem 1” and “Sys­tem 2”, I also cov­ered sub­agents in a more gen­eral way, as also in­clud­ing e.g. the kinds of sub­sys­tems which carry out ob­ject recog­ni­tion and are used to carry out tasks like ar­ith­metic. This was also the lens through which I looked at sub­agents in my sum­mary of Con­scious­ness and the Brain. Which kind of view is the most use­ful, de­pends on ex­actly what phe­nomenon we are try­ing to un­der­stand.