The Evolution of Sentience as a Factor in the Cambrian Explosion: Setting up the Question

Many thanks to Kieran Greig for fund­ing me to write this post and for feed­back on it. Thanks as well to Ja­son Schukraft and Daniela Wald­horn for feed­back on the post.

I use the term ‘sen­tience’ syn­ony­mously with ‘phe­nom­e­nal con­scious­ness.’ I use a defi­ni­tion by ex­am­ple – your ex­pe­rience of savour­ing a good meal or of stub­bing your toe are ex­am­ples of sen­tient ex­pe­rience. I’m prin­ci­pally con­cerned with valenced sen­tience and as a sim­plify­ing as­sump­tion I write as though sen­tience im­plies valenced sen­tience even though they are prob­a­bly sep­a­rable. I don’t have enough space in this post to fully flesh out my po­si­tion on con­scious­ness, but they are fairly similar to Luke Muehlhauser’s po­si­tion as ar­tic­u­lated in his Re­port on Con­scious­ness and Mo­ral Pa­tient­hood.

Con­text for this post

I have been re­search­ing in­ver­te­brate sen­tience for ap­prox­i­mately the past year and I be­lieve this post is rele­vant to that sub­ject. For more back­ground on the re­search that I’ve been do­ing look at this post here.

Think­ing about how and why sen­tience evolved is rele­vant for find­ing out which be­ings may be sen­tient. It may en­able us to say that most or all an­i­mals de­scend­ing from a cer­tain an­ces­tor are sen­tient. It’s par­tic­u­larly use­ful for cases we are most un­cer­tain about such as in­sects and other in­ver­te­brates.[i] This can bet­ter in­form de­ci­sions about helping an­i­mals. It can also in­form us about how con­tin­gent sen­tience may be, which can be rele­vant in as­sess­ing how likely it is that differ­ent types of digita or alien minds are sen­tient.[ii]

One ex­am­ple of how learn­ing about the evolu­tion of sen­tience may be rele­vant is the case of oc­to­puses and other cephalopods. There is some strong rea­sons to think that oc­to­puses may be con­scious. But cephalopods are very dis­tantly evolu­tion­ar­ily re­lated to us. They also likely evolved from quite sim­ple an­i­mals pos­si­bly re­sem­bling snails.

If we be­lieve oc­to­puses are sen­tient this leaves us with two main op­tions: be­lieve that sen­tience evolved at least twice, or be­lieve that our com­mon an­ces­tor with them was sen­tient. The former op­tion com­mits us to po­si­tion that prob­a­bly im­plies that alien life and to a lesser ex­tent digi­tal minds are more likely to be con­scious then we oth­er­wise would think. This is be­cause it in­di­cates that con­scious­ness is more eas­ily mul­ti­ply re­al­iz­able then we would oth­er­wise think. The lat­ter op­tion im­plies that at least some sim­ple minds are con­scious. Both of these po­ten­tial con­clu­sions are in­ter­est­ing and im­por­tant.

Limi­ta­tions of this post

I’m not an ex­pert on the sub­ject I’m writ­ing about in this post. This post is quite spec­u­la­tive and pre­limi­nary. Much of the post is set up, and I don’t reach any firm con­clu­sions. I may fol­low this post up with an­other post that gets deeper into the is­sue.

One fac­tor that calls into ques­tion some of the rea­son­ing in this post is some of the newer forms of ge­netic and molec­u­lar anal­y­sis on evolu­tion­ary his­tory that push back the com­mon an­ces­tors of many tax­o­nomic groups to much ear­lier than the stan­dard fos­sil driven nar­ra­tive would have it. If these are to be be­lieved then that pushes in favour of a con­clu­sion that the Cam­brian ex­plo­sion is not ex­cep­tional and so does not need an ex­cep­tional ex­pla­na­tion, such as be­ing driven by the evolu­tion of sen­tience.

In keep­ing with this, an­other ar­gu­ment that has been put for­ward is that the Cam­brian ex­plo­sion may ap­pear to be a greater break with evolu­tion­ary con­ti­nu­ity than it ap­pears to be only be­cause we lack as many fos­sils from ear­lier pe­ri­ods. The dis­cov­ery of the Edi­a­caran biota fills in this to some ex­tent and sup­ports the claim that the Cam­brian ex­plo­sion is less of a break with pre­vi­ous evolu­tion­ary con­ti­nu­ity then once be­lieved.

I will mostly be rea­son­ing us­ing the fos­sil driven model, be­cause that is the one that I know bet­ter, and I be­lieve the one that has been best de­vel­oped in the liter­a­ture, but be aware that this other ev­i­dence ex­ists that calls it into ques­tion.

The Explosion

The Cam­brian ex­plo­sion is a mys­tery of evolu­tion. It oc­curred in the pre­his­toric oceans from Per­sia around 545 to 520 mil­lion years ago. It is called an ‘ex­plo­sion’ be­cause of the vast num­ber of di­verse an­i­mal body plans that seem to sud­denly emerge out of nowhere. Many of these crea­tures were ex­tremely alien look­ing and difficult to clas­sify in ex­ist­ing tax­o­nomic group­ings. Stephen Jay Gould, an ad­vo­cate of the po­si­tion that evolu­tion of­ten pro­gresses through sud­den leaps and bounds, has cited it to sup­port his po­si­tion.

While ex­am­in­ing spec­i­mens from the Cam­brian ex­plo­sion Si­mon Con­way Mor­ri­son is re­ported to have said “Oh fuck, not an­other phy­lum!” More re­cent analy­ses have tended to push against the most rad­i­cal pic­tures of what the Cam­brian ex­plo­sion was, but it is still gen­er­ally agreed that some­thing like the Cam­brian ex­plo­sion took place and re­quires some sort of ex­pla­na­tion.

Many di­verse fac­tors have been posited as ex­pla­na­tions for the Cam­brian ex­plo­sion. Some of these are en­vi­ron­men­tal or ecolog­i­cal fac­tors such as an in­crease in the level of oxy­gen en­abling differ­ent body plans and oth­ers are evolu­tion­ary in­no­va­tions such as an­i­mal to an­i­mal pre­da­tion or dis­tance senses.

One thing that seems likely to me is that there was a cas­cade of differ­ent fac­tors. I be­lieve this be­cause many of the fea­tures that seem to have evolved seem like they would would in­evitably spur fu­ture evolu­tion and be­cause a cas­cade of differ­ent fac­tors is a more pow­er­ful ex­pla­na­tion than any sin­gle fac­tor. Here’s an ex­am­ple of how this might go: there may have been an in­crease in the amount of oxy­gen, which en­abled or­ganisms to have more com­plex and en­er­get­i­cally de­mand­ing bod­ies, which en­abled us the evolu­tion of dis­tance senses, which en­abled pre­da­tion, which caused se­lec­tion pres­sure for defences to pre­da­tion, and so on.

What is the func­tional role of sentience

Now in or­der to spec­u­late on sen­tience as a fac­tor in the Cam­brian ex­plo­sion we must spec­u­late on the func­tional role of sen­tience. The defi­ni­tion of sen­tience I gave ear­lier is in­tended to leave open what it’s func­tional role is, so this should not be taken as build­ing more into the defi­ni­tion, only as spec­u­la­tion on top of the already given defi­ni­tion.

To think about this ques­tion we must think about the func­tional roles that sen­tience might play. I’m not claiming that any en­tity that has ap­prox­i­mately these ca­pa­bil­ities is there­fore sen­tient. In­deed, I think it is prob­a­bly the case that all of these may some­times be car­ried out un­con­sciously in hu­mans. Some of the main roles that it seems to me that it plays are:

1) It al­lows for differ­ent types of re­in­force­ment learn­ing. The con­scious ex­pe­rience of pain or plea­sure re­spec­tively causes us to avoid or ap­proach phe­nom­ena or ac­tions as­so­ci­ated with that ex­pe­rience. Prob­a­bly some ba­sic level of as­so­ci­a­tive learn­ing oc­curs in some or­ganisms with­out sen­tience, but for at least more com­pli­cated forms of learn­ing on the ba­sis of ex­pe­rience, it is han­dled in hu­mans through us feel­ing pain and plea­sure.

2) It al­lows us to rank differ­ent al­ter­na­tives and weigh con­flict­ing op­tions against each other to give it an over­all ‘util­ity func­tion’ that al­lows us to de­cide on the best course of ac­tion. For ex­am­ple, if you de­cide to eat a pineap­ple you make that de­ci­sion on the ba­sis of var­i­ous plea­sures and pains as­so­ci­ated with that de­ci­sion and de­ter­mined that it is pos­i­tive on bal­ance.

3) It al­lows us to re­spond in the ap­pro­pri­ate way to cur­rent pos­i­tive and nega­tive stim­uli. For ex­am­ple, that al­lows us to avoid preda­tors and iden­tify and eat good food. Some of this pro­cess is han­dled through re­flexes in us, which ap­pear to be un­con­scious, but most of the pro­cess, es­pe­cially those parts in­volv­ing (2), are han­dled con­sciously in us.

4) It fa­cil­i­tates us hav­ing differ­ent con­scious emo­tions. Differ­ent moods or emo­tions are ex­pe­rienced con­sciously in us. Th­ese states have differ­ent re­spec­tive func­tional roles. The con­scious ex­pe­rience of the emo­tion seems to fa­cil­i­tate the differ­ent be­havi­our that differ­ent mood states tend to lead us to.

5) In ad­di­tion, non-af­fec­tive sen­tience plays the role of al­low­ing the or­ganism to sense ob­jects and have a spa­tial model of the world. This may be han­dled un­con­sciously in the case of blind­sight, so it may not nec­es­sar­ily be a con­scious pro­cess, though it is typ­i­cally han­dled con­sciously in hu­mans.

Of course, it could be the case that there is some ad­di­tional com­plex­ity on top of these fac­tors that needs to be added be­fore an or­ganism that can do these things can right­fully be called ‘sen­tient.’ This seems likely to me, though I think as the or­ganism comes to look in­creas­ingly like it is sen­tient, the like­li­hood that it is miss­ing whichever of these fac­tors that would make it ‘truly sen­tient’ de­creases.

I think that we should also re­sist the po­si­tion that there is some spe­cial sin­gle fac­tor that when added to an un­con­scious or­ganism sud­denly makes it con­scious. This is be­cause I think the ar­gu­ments for par­tic­u­lar pro­posed fea­tures for this such as mir­ror self-recog­ni­tion are poor and be­cause a the­ory such as this might be imag­in­ing some­thing similar to a Carte­sian the­atre pic­ture of con­scious­ness where some ‘magic spark’ is re­quired.

Did sen­tience help drive the Cam­brian ex­plo­sion?

If this anal­y­sis is cor­rect, sen­tience ap­pears to be par­tic­u­larly use­ful for ac­tive motile or­ganisms that need to avoid preda­tors. This ties in with some of my pre­vi­ously men­tioned pro­pos­als of what fac­tors drove the Cam­brian ex­plo­sion. For this rea­son, it seems to me like sen­tience may at least be part of the cas­cade of fac­tors that drove the Cam­brian ex­plo­sion.

It also seems to be the case that some of the an­i­mals that evolved in the Cam­brian ex­plo­sion are quite plau­si­bly sen­tient. For ex­am­ple, Ano­ma­lo­caris seems to have been the apex preda­tor of the time and was a me­tre long swim­ming arthro­pod with flex­ible ‘arms’ near its mouth and eyes ap­prox­i­mately as pow­er­ful as those of drag­on­flies. This sug­gests that at least in the per­haps ex­cep­tional case of Ano­ma­lo­caris, sen­tience did evolve in the Cam­brian ex­plo­sion or an ear­lier era.

Another Cam­brian an­i­mal is Haik­ouichthys, an early fish. Some au­thors have re­cently ar­gued to be per­haps the first sen­tient an­i­mal. Some of the rea­sons they give for this are it’s ad­e­quate image form­ing eyes an ad­di­tional senses that may have led to more com­plex neu­ral hi­er­ar­chies. It’s difficult to imag­ine ac­tive motile an­i­mals with dis­tance senses such as Haik­ouichthys that had to cope with preda­tors such as Ano­ma­lo­caris not hav­ing some­thing that at least re­sem­bles sen­tience. I have writ­ten an­other post here where I give more than rea­sons why I ex­pect large ac­tive an­i­mals such as anoma­lo­caris to be sen­tient.

How­ever, much of the di­ver­sifi­ca­tion that rep­re­sents the ex­plo­sion may have hap­pened in an­i­mals that were sim­ple enough that are more likely than not non-sen­tient. An­i­mal such as Ano­ma­lo­caris and Haik­ouichthys may be sen­tient, but most other an­i­mals, in­clud­ing the an­ces­tors of these an­i­mals, are more sim­ple and less likely to be con­scious.

How­ever, even if they didn’t them­selves rep­re­sent much of the di­ver­sity, arthro­pods have com­plex ac­tive bod­ies and may have driven di­ver­sifi­ca­tion in other groups. Other au­thors have sug­gested that pre­da­tion by arthro­pods may have been sig­nifi­cant in driv­ing the ex­plo­sion.

It’s difficult to tell what role sen­tience would’ve played in the Cam­brian ex­plo­sion. It could’ve been a pivotal fac­tor. Or it could have con­tributed rel­a­tively lit­tle and have been mostly a byproduct. It’s even more difficult to say if it would’ve been sen­tience or some sort of proto-sen­tience with many of the same func­tional prop­er­ties that I have sug­gested, but some­how not quite there to fully count as sen­tience and so not be morally rele­vant.

Si­mona Gins­berg and Eva Jablonka 2010 ar­gue that the evolu­tion of as­so­ci­a­tive learn­ing may have driven the Cam­brian ex­plo­sion. As­so­ci­a­tive learn­ing may be re­lated to sen­tience in cer­tain ways, but at least I’d sim­plest forms it does not seem to re­quire sen­tience. If they are right then per­haps sen­tience played rel­a­tively lit­tle role in ex­plo­sion and was merely ‘along for the ride’ be­cause as­so­ci­a­tive learn­ing would have prob­a­bly evolved ear­lier than sen­tience.[iii]

[i] It is par­tic­u­larly helpful with in­ver­te­brate con­scious­ness be­cause in­ver­te­brate con­scious­ness is a less tractable and straight­for­ward is­sue than ver­te­brate con­scious­ness, and so any tool that can help with the ques­tion is

[ii] One of the prin­ci­ple ways that study­ing the evolu­tion of sen­tience can help us with ques­tions about which digi­tal be­ings are sen­tient is by mov­ing us to­wards or away from a con­scious­ness is rare ver­sus con­scious­ness is com­mon con­clu­sion. In brief, if con­scious­ness evolves mul­ti­ple times that in­di­cates that it is more mul­ti­ple re­al­iz­able than we might oth­er­wise have ex­pected.

[iii] One over­sight in the Gins­burg pa­per is that they seem not to be aware of ev­i­dence of as­so­ci­a­tive learn­ing in plants and pro­tists. If pro­tists do show as­so­ci­a­tive learn­ing, this sug­gests that as­so­ci­a­tive learn­ing evolved much ear­lier and so could not ex­plain the Cam­brian ex­plo­sion. Of course, it could be that it was a mechanism that al­lowed for more com­plex as­so­ci­a­tive learn­ing then we see in pro­tists that evolved later on and drove the Cam­brian ex­plo­sion. Sen­tience could also have al­lowed for this higher level of as­so­ci­a­tive learn­ing, but this is pure spec­u­la­tion.