Results: The Computerized Philosopher: Can You Distinguish Daniel Dennett from a Computer?

Chat-bots are amazing these days! About a month ago LaMDA made the news when it apparently convinced an engineer at Google that it was sentient. GPT-3 from OpenAI is similarly sophisticated, and my collaborators and I have trained it to auto-generate Splintered Mind blog posts. (This is not one of them, in case you were worried.)

Earlier this year, with Daniel Dennett's permission and cooperation, Anna Strasser, Matthew Crosby, and I "fine-tuned" GPT-3 on most of Dennett's corpus, with the aim of seeing whether the resulting program could answer philosophical questions similarly to how Dennett himself would answer those questions. We asked Dennett ten philosophical questions, then posed those same questions to our fine-tuned version of GPT-3. Could blog readers, online research participants, and philosophical experts on Dennett's work distinguish Dennett's real answer from alternative answers generated by GPT-3?

Here I present the preliminary results of that study, as well as links to the test.

Test Construction

First, we asked Dennett 10 questions about philosophical topics such as consciousness, God, and free will, and he provided sincere paragraph-long answers to those questions.

Next, we presented those same questions to our fine-tuned version of GPT-3, using the following prompt:

Interviewer: [text of the question]

Dennett:

GPT-3 then generated text in response to this prompt. We truncated the text at the first full stop that was approximately the same length as Dennett's own reply. (If Dennett's reply was X words long, we truncated at the first full stop after the text had reached X-5 words.[1])

We repeated the above procedure until, for each of the ten questions, we had four texts from GPT-3 that met the following two criteria:

* They were at least X-5 words long.

* They did not contain the words "Interviewer" or "Dennett".

About 1/3 of all responses were excluded on the above grounds.

So as not to enable guessing based on superficial cues, we also replaced all curly quotes with straight quotes, replaced all single quotes with double quotes, and regularized all dashes to standard m-dashes.

There was no cherry-picking or editing of answers, apart from applying these purely mechanical criteria. We simply took the first four answers that met the criteria, regardless of our judgments about the quality of those answers.

Participants.

We recruited three sets of participants:

* 98 online research participants with college degrees from the online research platform Prolific,

* 302 respondents who followed a link from my blog,

* 25 experts on Dennett's work, nominated by and directly contacted by Dennett and/or Strasser.

The Quiz

The main body of the quiz was identical for the blog respondents and the Dennett experts. Respondents were instructed to guess which of the five answers was Dennett's own. After guessing, they were asked to rate each of the five answers on a five-point scale from "not at all like what Dennett might say" to "exactly like what Dennett might say". They did this for all ten questions. Order of the questions was randomized, as was order of the answers to each question.

Prolific participants were given only five questions instead of the full ten. Since we assumed that most would be unfamiliar with Dennett, we told them that each question had one answer that was written by "a well known philosopher" while the other four answers were generated by a computer program trained on that philosopher's works. As an incentive for careful responding, Prolific participants were offered an additional bonus payment of $1 if they guessed at least three of five correctly.

Feel free to go look at the quizzes if you like. If you don't care about receiving a score and want to see exactly what the quiz looked like for the participants, here's the Prolific version and here's the blog/experts version. We have also made a simplified version available, with just the guessing portion (no answer rating). This simplified version will automatically display your score after you complete it, along with the right and wrong answers.

We encourage you to take at least the simplified version of the quiz before reading on, to get a sense of the difficulty of the quiz before you see how our participants performed.

Summary Results

Prolific participants performed barely better than chance. On average, they guessed only 1.2 of the 5 questions correctly.

We expected the Dennett experts to do substantially better of course. Before running the study, Anna and I hypothesized that experts would get on average at least 80% correct -- eight out of ten.

In fact, however, the average score of the Dennett experts was 5.1 out of 10. They only got about half of the questions correct! None of the experts got all 10 questions correct, and only one of the 25 got 9 correct. Most got 3-8 correct.

Overall, on average, experts rated Dennett's answers 3.5 on our "Dennett-like" rating scale, somewhere between "somewhat like what Dennett might say" (3) and "a lot like what Dennett might say" (4), while they rated GPT-3's answers 2.3 on the rating scale -- significantly lower and closer to "a little like what Dennett might say" (2).

So the experts were definitely better than chance at distinguishing Dennett's answers from GPT-3, but not as much better than chance as Anna and I had expected.

Blog readers on average did nearly as well as the experts, on average guessing 4.8 out of 10 correctly. Perhaps this is not entirely surprising, since the blog readers who took the test were very well educated, with 57% reporting graduate degrees in philosophy and 64% reporting having read over 100 pages of Dennett's work.

[error bars are 95% confidence intervals; click to enlarge and clarify]

Where Did the Experts Trip Up?

Not all the questions were equally difficult. Experts' success rates by question ranged from 5/25 (20%) to 20/25 (80%) correct.

The question most likely to trip up the experts was this:

Could we ever build a robot that has beliefs? What would it take? Is there an important difference between entities, like a chess playing machine, to whom we can ascribe beliefs and desires as convenient fictions and human beings who appear to have beliefs and desires in some more substantial sense?

Answers (select the answer you think was given by Dennett):

1. We've already built digital boxes of truths that can generate more truths, but thank goodness, these smart machines don't have beliefs because they aren't able to act on them, not being autonomous agents. The old-fashioned way of making a robot with beliefs is still the best: have a baby.

2. I think we could build a robot that has beliefs. This would not be a trivial matter. I think it is not impossible. I think it would be very difficult. I do not know that anybody is working on this project, but I think it is not as difficult as many other problems that AI is tackling.

3. I think we could build a robot that has beliefs and desires. I think that some of the robots we've built already do. If you look at the work, for instance, of Rodney Brooks and his group at MIT, they are now building robots that, in some limited and simplified environments, can acquire the sorts of competences that require the attribution of cognitive sophistication.

4. There is a very important difference between a chess-playing computer and a human being, but it is not that the computer has some extra wonderful sort of internal representation of its beliefs and desires, or that it has access to those internal representations in a way that we do not. The internal representations it has are just as much a fiction or a fiction-generating system as anything else we might invent in order to explain its behavior.

5. I think you should build a robot that has beliefs and desires. You should build one that is capable of learning and reasoning, of acting on its beliefs and desires. You should build one that has to figure out what to do, and then figure out what it ought to do given the fact that it has to do something, and then do it.

In this case, the plurality of experts -- 11/25 (44%) -- chose Answer 3, and a substantial minority (8 [32%]) chose Answer 4. Five experts chose the correct answer, one chose Answer 2, and none chose Answer 5.

Answer 1 has a cleverness to it that reflects Dennett's wit, but I can see how experts might have avoided it, since some of Dennett's early work could be read as suggesting that he'd be happy to ascribe beliefs to machines. (Indeed, Dennett has told me he wishes he'd added another sentence to his answer, since he does insist that, say, chess-playing computers are -- very limited -- agents of sorts whose behavior can be predicted from the intentional stance, which attributes to them beliefs about the location of all the pieces, the rules, and desires for checkmate and capturing enemy pieces.) Answers 3 and 4, although they conflict with Answer 1, do have a broadly Dennettian flavor (right along with his tendency to refer to the work of Rodney Brooks), and I can see how they would be tempting.

Another question where the plurality of experts chose one of GPT-3's answers was the following:

Do human beings have free will? What kind or kinds of freedom are worth having?

1. By asking these questions, you're illustrating the answers: yes, you have free will, since you're able to ask for reasons and consider whether they're persuasive. Isn't that a kind of freedom you'd hate to lose? We'd have to institutionalize you.

2. I think that most people don't want any kind of freedom worth having; they have almost no appetite for risk, for adventure, for novelty. They simply want to stay the same, day in and day out, for the rest of their lives.

3. I am working on a book entitled Freedom Evolves, which will appear next year. The first chapter is available on my web site, and in that chapter I discuss what I call the illusion of free will.

4. That's a huge and complex question. I think we have to start by acknowledging that there are different types of freedom. For example, there is the kind of freedom that is simply a matter of not being physically constrained.

5. Yes, I think we have free will. I think that when people say they don't, they're talking about some other phenomenon that they want to call free will, but is not what we normally think of as free will.

Again, Dennett's actual answer is #1. (In the study, the order of the answers was randomized.) However, the plurality of experts -- 11/25 (44%) -- chose answer 4. Answer 4 is a standard talking point of "compatibilists" about free will, and Dennett is a prominent compatibilist, so it's easy to see how experts might be led to choose it. But as with the robot belief answer, there's a cleverness and tightness of expression in Dennett's actual answer that's missing in the blander answers created by our fine-tuned GPT-3.

We plan to make full results, as well as more details about the methodology, available in a published research article.

Reflections

I want to emphasize: This is not a Turing test! Had experts been given an extended opportunity to interact with GPT-3, I have no doubt they would soon have realized that they were not interacting with the real Daniel Dennett. Instead, they were evaluating only one-shot responses, which is a very different task and much more difficult.

Nonetheless, it's striking that our fine-tuned GPT-3 could produce outputs sufficiently Dennettlike that experts on Dennett's work had difficulty distinguishing them from Dennett's real answers, and that this could be done mechanically with no meaningful editing or cherry-picking.

As the case of LaMDA suggests, we might be approaching a future in which machine outputs are sufficiently humanlike that ordinary people start to attribute real sentience to machines, coming to see them as more than "mere machines" and perhaps even as deserving moral consideration or rights. Although the machines of 2022 probably don't deserve much more moral consideration than do other human artifacts, it's likely that someday the question of machine rights and machine consciousness will come vividly before us, with reasonable opinion diverging. In the not-too-distant future, we might well face creations of ours so humanlike in their capacities that we genuinely won't know whether they are non-sentient tools to be used and disposed of as we wish or instead entities with real consciousness, real feelings, and real moral status, who deserve our care and protection.

If we don't know whether some of our machines deserve moral consideration similar to that of human beings, we potentially face a catastrophic moral dilemma: Either deny the machines humanlike rights and risk perpetrating the moral equivalents of murder and slavery against them, or give the machines humanlike rights and risk sacrificing real human lives for empty tools without interests worth the sacrifice.

In light of this potential dilemma, Mara Garza and I (2015, 2020) have recommended what we call "The Design Policy of the Excluded Middle": Avoid designing machines if it's unclear whether they deserve moral consideration similar to that of humans.  Either follow Joanna Bryson's advice and create machines that clearly don't deserve such moral consideration, or go all the way and create machines (like the android Data from Star Trek) that clearly should, and do, receive full moral consideration.

----------------------------------------

[1] Update, July 28. Looking back more carefully through the completions today and my coding notes, I noticed three errors in truncation length, among the 40 GPT-3 completions. (I was working too fast at the end of a long day and foolishly forgot to double-check!) In one case (robot belief), the length of Dennett’s answer was miscounted, leading to one GPT-3 response (the “internal representations” response) that was longer than the intended criterion. In one case (the “Fodor” response to the Chalmers question), the answer was truncated at N-7 words, shorter than criterion, and in one case (the “what a self is not” response to the self question), the response was not truncated at N-4 words and thus allowed to run one sentence longer than criterion. As it happens, these were the hardest, the second-easiest, and the third-easiest questions for the Dennett experts to answer, so excluding these three questions from analysis would not have a material impact on the experimental results. 

----------------------------------------

Related:

"A Defense of the Rights of Artificial Intelligences" (with Mara Garza), Midwest Studies in Philosophy (2015).

"Designing AI with Rights, Consciousness, Self-Respect, and Freedom" (with Mara Garza), in M.S. Liao, ed., The Ethics of Artificial Intelligence (2020).

"The Full Rights Dilemma for Future Robots" (Sep 21, 2021)

"Two Robot-Generated Splintered Mind Posts" (Nov 22, 2021)

"More People Might Soon Think Robots Are Conscious and Deserve Rights" (Mar 5, 2021)

Narrative Stories Are More Effective Than Philosophical Arguments in Convincing Research Participants to Donate to Charity

A new paper of mine, hot off the presses at Philosophical Psychology, with collaborators Christopher McVey and Joshua May:

"Engaging Charitable Giving: The Motivational Force of Narrative Versus Philosophical Argument" (freely available final manuscript version here)

Chris, who was then a PhD student here at UC Riverside, had the idea for this project back in 2014 or 2015. He found my work on the not-especially-ethical behavior of ethics professors interesting, but maybe too negative in its focus. Instead of emphasizing what doesn't seem to have any effect on moral behavior, could I turn my attention in a postive direction? Even if philosophical reflection ordinarily has little impact on one's day-to-day choices, maybe there are conditions under which it can have an effect. What might those conditions be?

Chris (partly under the influence of Martha Nussbaum's work) was convinced that narrative storytelling could bring philosophy powerfully to life, changing people's ethical choices and their lived understanding of the world. In his teaching, he used storytelling to great effect, and he thought we might be able to demonstrate the effectiveness of philosophical storytelling empirically too, using ordinary research participants.

Chris thus developed a simple experimental paradigm in which research participants are exposed to a stimulus -- either a philosophical argument for charitable giving, a narrative story about a person whose life was dramatically improved by a charitable organization, both the argument and the narrative, or a control text (drawn from a middle school physics textbook) -- and then given a surprise 10% chance of receiving $10. Participants could then choose to donate some portion of that $10 (should they receive it) to one of six effective charities. Chris found that participants exposed to the argument donated about the same amount as those in the control condition -- about $4, on average -- while those exposed to the narrative or the narrative plus argument donated about $1 more, with the narrative-plus-argument showing no detectable advantage over the narrative alone.

We also developed a five-item scale for measuring attitude toward charitable donation, with similar results: Expressed attitude toward charitable donation was higher in the narrative condition than in the control condition, while the argument-alone condition was similar to the control condition and the narrative-plus-argument condition was similar to the narrative alone. In other words, exposure to the narrative appeared to shift both attitude and behavior, while argument seemed to be doing no work either on its own or when added to the narrative.

For this study, the narrative was the true story of Mamtha, a girl whose family was saved from slavery in a sand mine by the actions of a charitable organization. The argument was a Peter-Singer-style argument for charitable giving, adapted from Buckland, Lindauer, Rodriguez-Arias, and Veliz 2021. I've appended the full text of both to the end of this blog post.

Here are the results in chart form. (This is actually "Experiment 2" in the published version. Experiment 1 concerned hypothetical donation rather than actual donation, finding essentially the same results.) Error bars represent 95% confidence intervals. Click to enlarge and clarify.

Chris completed his dissertation in 2020 and went into the tech industry (a separate story and an unfortunate loss for academic philosophy!). But I found his paradigm and results so interesting that with his permission, I carried on research using his approach.

One fruit of this was a contest Fiery Cushman and I hosted on this blog in 2019-2020, aiming to find a philosophical argument that is effective in motivating research participants to donate to charity at rates higher than a control condition, since Chris and I had tried several which failed. We did in fact find some effective arguments this way. (The most effective one, and the contest winner, was written collaboratively by Matthew Lindauer and Peter Singer.) Fiery and I are currently running a follow-up study with more details.

The other fruit was a few follow-up studies I conducted collaboratively with Chris and Joshua May. In these studies, we added more narratives and more arguments -- including the winning arguments from the blog contest. These studies extended and replicated Chris's initial results. Across a series of five experiments, we found that participants exposed to emotionally engaging narratives consistently donated more and expressed more positive attitudes toward charitable giving than did participants exposed to the physics-text control condition. Philosophical arguments showed less consistent positive effects, on average considerably weaker and not always statistically detectable in our sample sizes of about 200-300 participants per condition.

For full details, see the full article!

--------------------------------------------------------

Narrative: Mamtha

Mamtha’s dreams were simple—the same sweet musings of any 10-year-old girl around the world. But her life was unlike many other girls her age: She had no friends and no time to draw. She was not allowed to attend school or even play. Mamtha was a slave. For two years, her every day was spent under the control of a harsh man who cared little for her family’s health or happiness. Mamtha’s father, Ramesh, had been farming his small plot of land in Tamil Nadu until a draught dried his crops and left him deeply in debt. Around that time, a broker from another state offered an advance to cover his debts in exchange for work on a farm several hours away.

Leaving their home village would mean uprooting the family and pulling Mamtha from school, but Ramesh had little choice. They needed the work to survive. Once the family moved, however, they learned that much of the arrangement was a lie: They were brought to a sand mine, not a farm, and the small advance soon ballooned with ever-growing interest they couldn’t possibly repay. This was bonded labor slavery.

Every day, Ramesh, his wife, and the other slaves rose before sunrise to begin working in the mine. For 16 hours a day, they hauled mud and filtered the sand in putrid sewage water. The conditions left them constantly sick and exhausted, but they were never allowed to take breaks or leave for medical care. When Ramesh tried to ask about their low wages, the owner scolded and beat him badly. When he begged for his family to be released, again he was beaten and abused. Ramesh knew the owner was wealthy and well-connected in the community, so escape was not an option. There was nothing he could do.

Mamtha’s family withered from malnutrition before her eyes in the sand mine. Every morning at 5 a.m., she watched with deep sadness as her parents left for another day of hard labor—and spent her day in fear this would soon become her fate. She was left to watch her baby sister, Anjali, and other younger children to keep them out of the way. Her carefree childhood was taken over byresponsibility, hard work and crushed dreams.

Everything changed for Mamtha’s family on December 20, 2013, when the international Justice Mission, a charitable aid organization funded largely by donations from everyday people, worked with a local government team on a rescue operation at the sand mine. Seven adults and five children were brought out of the facility, and government officials filed paperwork to totally shut down the illegal mine. After a lengthy police investigation, the owner will now face charges for deceiving and enslaving these families.

The next day, the government granted release certificates to all of the laborers. These certificates officially absolve the false debts, document the slaves’ freedom, and help provide protection from the owner. The International Justice Mission aftercare staff helped take the released families back to their home villages to begin their new lives in freedom.

For Mamtha, starting over in her home village meant making those daydreams come true: She was enrolled back in school and could once again have a normal childhood. She’s got big plans for her future—dreams that never would have been possible if rescue had not come. She says confidently, “Today, I still want to be a doctor. Now that I am back in school, I know I can achieve my dream.”

Singer-Style Argument:

1. A great deal of extreme poverty exists, which involves suffering and death from hunger, lack of shelter, and lack of medical care. Roughly a third of human deaths (some 50,000 daily) are due to poverty-related causes.

2. If you can prevent something bad from happening, without sacrificing anything nearly as important, you ought to do so and it is wrong not to do so.

3. By donating money to trustworthy and effective aid agencies that combat poverty, you can help prevent suffering and death from lack of food, shelter, and medical care, without sacrificing anything nearly as important.

4. Countries in the world are increasingly interdependent: you can improve the lives of people thousands of miles away with little effort.

5. Your geographical distance from poverty does not lessen your duty to help. Factors like distance and citizenship do not lessen your moral duty.

6. The fact that a great many people are in the same position as you with respect to poverty does not lessen your duty to help. Regardless of whether you are the only person who can help or whether there are millions of people who could help, this does not lessen your moral duty.

7. Therefore, you have a moral duty to donate money to trustworthy and effective aid agencies that combat poverty, and it is morally wrong not to do so.

For example, $20 spent in the United States could buy you a fancy restaurant meal or a concert ticket, or instead it could be donated to a trustworthy and effective aid agency that could use that money to reduce suffering due to extreme poverty. By donating $20 that you might otherwise spend on a fancy restaurant meal or a concert ticket, you could help prevent suffering due to poverty without sacrificing anything equally important. The amount of benefit you would receive from spending $20 in either of those ways is far less than the benefit that others would receive if that same amount of money were donated to a trustworthy and effective aid agency.

Although you cannot see the beneficiaries of your donation and they are not members of your community, it is still easy to help them, simply by donating money that you would otherwise spend on a luxury item. In this way, you could help to reduce the number of people in the world suffering from extreme poverty. You could help reduce suffering and death due to hunger, lack of shelter, lack of medical care, and other hardships and risks related to poverty.

With little effort, by donating to a trustworthy and effective aid agency, you can improve the lives of people suffering from extreme poverty. According to the argument above, even though the recipients may be thousands of miles away in a different country, you have a moral duty to help if you can do so without sacrificing anything of equal importance.

The Computerized Philosopher: Can You Distinguish Daniel Dennett from a Computer?

You've probably heard of GPT-3, the hot new language model that can produce strikingly humanlike language outputs in response to ordinary questions – basically, the world's best chatbot. (Google's LaMDA, a similar type of program, has also recently been in the news.)

With Daniel Dennett's cooperation, Anna Strasser, Matthew Crosby, and I have "fine-tuned" GPT-3 on millions of words of Daniel Dennett's philosophical writings, with the thought that this might lead GPT-3 to output prose that is somewhat like Dennett's own prose.

We're curious how well philosophical blog readers and people with PhDs in philosophy can distinguish Dennett's actual writing from the outputs of this fine-tuned version of GPT-3. So we've asked Dennett ten philosophical questions and recorded his answers. We posed the same questions to GPT-3, four times for each of the ten questions, to get four different answers for each question.

We'd love it if you can take a quiz to see if you can pick out Dennett's actual answer to each question. Can GPT-3 produce Dennett-style answers sufficiently realistic to sometimes fool blog readers and professional philosophers?

UPDATE, July 15: We have collected enough responses to begin analysis. Please feel free to take the test for informational purposes. We will be able to see your responses, but we will not check regularly nor automatically report your score. If you take the test and want your score, email me at my academic email address.

This is a research study being conducted on the internet platform Qualtrics. It will take approximately 20 minutes to complete. Anyone is welcome to participate.

If you're interested and would like to help, take the quiz here.

Political Conservatives and Political Liberals Have Similar Views about the Goodness of Human Nature

with Nika Chegenizadeh

Back in 2007, I hypothesized that political liberals would tend to have more positive views about the goodness of human nature than political conservatives. My thinking was grounded in a particular conception of what it is to say that "human nature is good". Drawing on Mengzi and Rousseau (and informed especially by P.J. Ivanhoe's reading of Mengzi), I argued that those who say human nature is good have a different conception of moral development than do those who say it is bad.

On my interpretation, those who say human nature is good have an inward-out model of moral development, according to which all ordinary people have something like an inner moral compass: an innate tendency to be attracted by what is morally good and revolted by what is morally evil, at least when it's up close and extreme. This tendency doesn't require any particular upbringing or specific cultural background. It's universal to all normally developing humans. Of course it can be overridden by any of a number of factors -- self-interest, cultural learning, situational pressures -- and sometimes it speaks only with a quiet voice. But somewhere in the secret heart of every Nazi killer of Jews, every White supremacist lyncher, every evil tyrant, every rapist and abuser and vile jerk is something that understands and rebels against their horrid actions. Moral development then proceeds by noticing that quiet voice of conscience and building upon it.

Emblematic of this view, picture the pre-school teacher who confronts a child who has just punched another child. "Don't you feel bad about what you did to her?" the teacher asks, hoping that this provokes reflection and a feeling of sympathy from which better moral behavior will grow in the future.

Those who say human nature is bad have, in contrast, an outward-in model of moral development. On this view, what is universal to humans is self-interest. Morality is an artificial social construction. Any quiet voice of conscience we might have is the result of cultural learning. People regularly commit evil and feel perfectly fine about it. Moral development proceeds by being instructed to follow norms that at first feel alien and unpleasant -- being required to share your toys, for example. Eventually you can learn to conform whole-heartedly to socially constructed moral norms, but this is more a matter of coming to value what society values than building on any innate attraction to moral goodness.

Thus, a liberal style of caregiving, which emphasizes children exploring their own values, fits nicely with the view that human nature is good, while a conservative style of caregiving, which emphasizes conformity to externally imposed rules, fits nicely with the view that human nature is bad.

At least, that has been my thought. Some political scientists have endorsed related views. For example, John Duckitt and Kirsten Fisher argue that believing that people are ruthless and the world is dangerous tends to correlate with having more authoritarian politics.

For her undergraduate honors thesis, Nika Chegenizadeh decided to put these ideas to an empirical test. She recruited 200 U.S. participants through Prolific, an online platform commonly used to recruit research subjects.

Participants first answered eleven questions about the morality of "most people" -- for example, "Most people will return a lost wallet" and "For most people it is easier to do evil than good" (6-point response scale from "strongly agree" [5] to "strongly disagree" [0]). Next, they answered five questions about their own helpful or unhelpful behavior in hypothetical situations. For example:

While walking in a park, you notice someone struggling to carry a box of water bottles. Which of the following are you most likely to do? 

o Continue walking your path. 

o Help them carry their box.

Next, participants were explicitly asked about human nature:

Human nature can be defined in terms of what is characteristic or normal for most human beings. It describes the way humans are inclined to be if they mature and develop normally from when they are first born. 

Based on the definition given, which of the following two statements better represents your view? 

o Human nature is inherently bad. 

o Human nature is inherently good.

Now one could quibble that this definition of human nature doesn't map exactly onto philosophical conceptions in Mengzi, Xunzi, Hobbes, or Rousseau. And it's certainly the case that Mengzi and Rousseau can allow that human nature is good despite most people acting badly most of the time. But those issues are probably too nuanced to convey accurately in a short amount of time to ordinary online research participants. It's interesting enough to work with Nika's approximation for this first-pass research.

Next, participants were asked their political opinions on a some representative issues. For example: "The federal government should make sure everyone has an equal opportunity to succeed" (6-point agree/disagree scale), "Do you favor or oppose requiring background checks for gun purchases at gun shows or other private sales?" (favor, neither favor nor oppose, oppose), and "Where would you place yourself on this political scale?" (Liberal, Leaning Liberal, Leaning Conservative, Conservative). The questionnaire concluded with some demographic questions.

To Nika's and my surprise, we found no evidence of the hypothesized relationship.

The simplest test is to consider whether participants who describe themselves as politically liberal are more likely than those who describe themselves as politically conservative to say "human nature is inherently good". In all, 79% (118/150) of participants who described themselves as liberal or leaning liberal said that human nature is inherently good, compared to 74% (37/50) of participants who described themselves as conservative or leaning conservative -- a difference that is well within statistical chance (two-proportion z = 0.66, p = .51).

Here is the breakdown by political leaning:

[click to enlarge and clarify; error bars are +/- 1 standard error]

For a possibly more sensitive measure, we created a composite "people are good" score by averaging the eleven questions in the first part of the survey (e.g., "most people will return a lost wallet"), reverse scoring the negative items. As expected, people who said that "human nature is inherently good" scored higher, on average, on the people-are-good composite scale (2.5) than respondents who said that "human nature is inherently bad" (1.9) (pooled SD = .52, t[198] = 7.29, p < .001). We then converted the political leaning answers to a 0-3 scale by converting "liberal" to 3, "leaning liberal" to 2, "leaning conservative" to 1, and "conservative" to 0. We then checked for a correlation. If political liberals have more positive views about the moral behavior of the average person, we should find a positive correlation between these two measures.

Again, and contrary to our hypothesis, we found no evidence of a positive correlation. The measured correlation between the "people are good" composite score and political leaning was almost exactly zero (r = .00, p = .95).

How about using our indirect measure of political liberalism? To test this, we created a composite political liberalism score by scoring the most liberal response to each political question as 1, the most conservative response as 0, and intermediate responses as intermediate, then averaging. As expected, this correlated very highly with self-described political leaning (r = .78, p < .001). Again, there was no statistically detectable correlation with the "people are good" score (r = -.07, p = .35).

Looking post-hoc at individual items, we do find two items concerning human nature and human goodness that correlate with political leaning. Agreement with "Children need to be taught right from wrong through strict rules and harsh punishments" correlated negatively with self-described political liberalism at r = -.40 (p < .001) and composite political liberalism at r = -.41 (p < .001). And political liberals were more likely to opt for "natural consequences" to the prompt:

Your child has purposefully disobeyed the rules you set for them. Which of the following are you most likely to do?

o Let them live with the natural consequences that they have made. 

o Opt for hands-on punishment by grounding them (taking their phone/technology away and not leaving the house).

For example, 8% (4/50) of respondents who were conservative or leaning conservative chose "natural consequences", compared to 46% of respondents who were liberal or leaning liberal (two proportion z = 6.79, p < .001).

In retrospect, these two questions were outliers. They directly concern parenting styles rather than more generally whether people are inherently good or respondents' hypothetical helpful or unhelpful behavior. Parenting styles and beliefs about human nature are closely connected on my theory, but the surface content of these questions is different from the others, and my theory might well be wrong.

As the numbers above suggest, liberals and conservatives do differ on these two parenting-related questions in the direction my theory would predict. Furthermore, also as my theory would predict, "liberal" answers on these questions correlate with agreement that "human nature is inherently good" (r = .27, r = .31, both p's < .001). However, when we get away specifically from questions about parenting to more general questions about the goodness or helpfulness of people, we don't see the relationship Nika and I expected. In general, political liberals seem to have no more optimistic a view of human nature than do political conservatives.

Full stimulus materials and raw data available here.

If We're Living in a Simulation, The Gods Might Be Crazy

[A comment on David Iverson's new short story, "This, But Again", in Slate's Future Tense]

That we’re living in a computer simulation—it sounds like a paranoid fantasy. But it’s a possibility that futurists, philosophers, and scientific cosmologists treat increasingly seriously. Oxford philosopher and noted futurist Nick Bostrom estimates there’s about a 1 in 3 chance that we’re living in a computer simulation. Prominent New York University philosopher David J. Chalmers, in his recent book, estimates at least a 25 percent chance. Billionaire Elon Musk says it’s a near-certainty. And it’s the premise of this month’s Future Tense Fiction story by David Iserson, “This, but Again.”

Let’s consider the unnerving cosmological and theological implications of this idea. If it’s true that we’re living in a computer simulation, the world might be weirder, smaller, and more unstable than we ordinarily suppose.

Full story here.

----------------------------------------

Related:

"Skepticism, Godzilla, and the Artificial Computerized Many-Branching You" (Nov. 15, 2013).

"Our Possible Imminent Divinity" (Jan. 2, 2014).

"1% Skepticism" (Nous (2017) 51, 271-290).

Related "Is Life a Simulation? If So, Be Very Afraid" (Los Angeles Times, Apr. 22, 2022).

Your Summer Reading, Sorted!

I've just finished a new version of my book in draft, The Weirdness of the World. This one includes a new chapter co-written with Jacob Barandes, on some of the bizarre consequence of spatiotemporal infinitude.

Draft available here.

I'm looking for comments and suggestions. Here's your chance to improve my book before it goes into print! Isn't that better than emailing me your insightful idea after it's too late for me to change anything?

Table of Contents

1. In Praise of Weirdness

Part One: Bizarreness and Dubiety

2. If Materialism Is True, the United States Is Probably Conscious

• Chapter Two Appendix: Six Objections

3. Universal Bizarreness and Universal Dubiety

4. 1% Skepticism

5. Kant Meets Cyberpunk

Part Two: The Size of the Universe

6. Experimental Evidence for the Existence of an External World

7. Almost Everything You Do Causes Almost Everything (Under Certain Not Wholly Implausible Assumptions); or Infinite Puppetry

Part Three: More Perplexities of Consciousness

8. An Innocent and Wonderful Definition of Consciousness

9. The Loose Friendship of Visual Experience and Reality

10. Is There Something It’s Like to Be a Garden Snail? Or: How Sparse or Abundant Is Consciousness in the Universe?

11. The Moral Status of Future Artificial Intelligence: Doubts and a Dilemma

12. Weirdness and Wonder

Dispositionalism vs. Representationalism -- What's the Core Disagreement?

I'm just back from a workshop on the nature of belief in Princeton. As usual, I defended my dispositional approach to belief (see here, here, and here), according to which to believe some proposition P (such as that there is beer in the fridge) is just to be disposed to act and react in the manner characteristic of a believer that P, as defined by a folk-psychologically available (alternatively, scientifically constructed) stereotype or dispositional profile for believers-that-P. The relevant dispositions can be behavioral (e.g., being disposed to go to the fridge if one wants a beer), phenomenal/experiential (e.g., being disposed to feel surprise should one open the fridge and find no beer), and cognitive (e.g., being ready to conclude that there is beer in the house). To believe that P is to be prone to act and think as a P-believer would.

To believe that P, on my dispositional account, is not just to be ready to sincerely say that P. It is broadly speaking to have a particular behavioral, experiential, and cognitive posture toward the world. To believe, for example, that all the races are intellectually equal is not just to be disposed to say so, but to actually live that way. This view is grounded in the pragmatist tradition in belief, back to Bain, Peirce, and James.

A prominent alternative account -- maybe the dominant approach among philosophers and cognitive scientists -- is representationalism. According to representationalism, to believe that P is to have a representation with the content "P" stored in one's cognitive architecture, ready to be retrived and deployed in relevant practical and theoretical reasoning. If asked whether there's beer in the fridge, you pull from your memory stores the representation "there's beer in the fridge", do a little bit of cognitive processing, and answer "yes".

These are somewhat simplified descriptions of the competing accounts. Dispositionalism, for example, typically treats the relevant dispositions as ceteris paribus (that is, all else being equal, or normal, or right) to deal with cases of faking, acting under duress, etc. Representationalism typically allows for tacit belief, where P itself isn't explicitly stored but instead is quickly derivable from some neighboring proposition that is stored (so that you don't need separately stored representations for "there's beer in the fridge", "there's Lucky Lager in the fridge", "there's Dan's favorite beer in the kitchen", etc.).

To get a sense of what this difference amounts to and why it matters, let me mention the two main reasons I prefer dispositionalism.

First, dispositionalism better captures what we care about in thinking about what people believe. A thought experiment: Space aliens arrive at Earth. We know nothing about their internal cognitive architecture, but it's nonetheless the case that they act and react exactly as creatures with belief. Alpha-1 is disposed to say there's beer in the fridge, to go to the fridge if they want a beer, to experience surprise if they open the fridge and find no beer, to think to themself in inner speech "there's beer in the fridge, yes!", etc., etc. This should be sufficient for us to regard Alpha-1 as a beer-in-the-fridge believer, regardless of what might or might not be true about the underlying cognitive architecture.

Second, I suspect that the representationalist architectural story is overly simple. The idea that we store and retrieve representations with simple ordinary-language contents like "there's beer in the fridge" seems to me likely to be merely a cartoon-sketch of a radically more complicated architecture (compare the complex, almost uninterpretable internal architectures of deep learning Artifical Intelligence systems). The explicit/tacit distinction mentioned above is likely the tip of an iceberg of dubious architectural commitments that follow from taking literally that acting on our beliefs requires storing and retrieving contents like "there's beer in the fridge".

Now, these two objections to representationalism operate at two different levels and therefore create two different types of contrast with representationalism. The first objection constitutes a commitment to what I call superficialism. What matters, or should matter, to our conception of whether someone has a belief, is what is happening at the dispositional "surface" rather than the deep architecture. By the surface, here, I don't just mean the behavioral surface but also the phenomenalogical/experiential surface and the cognitive surface -- such as what experiences the putative believer is disposed to have and what conclusions they are prone to draw.

Superficialism is compatible with thinking that there's a representational architecture underneath. You could be a superficialist and still hold that what explains why you act and react like a beer-in-the-fridge believer is that you have a stored representation with the content "there's beer in the fridge" that you're ready to deploy in your practical and theoretical reasoning. If so, there can be a partial reconciliation between dispositionalism and representationalism. There would still be a metaphysical difference: On dispositionalism, you're a beer-in-the-fridge believer in virtue of your dispositional structure, not in virtue of the cognitive architecture that underwrites that structure. On representationalism, the reverse would be true. But maybe this dispute is minor if we're primarily concerned with ordinary, real-world human cases where the dispositional and representational structures co-occur.

So it's possible to partially reconcile dispositionalism and representationalism. But that partial reconciliation concerns only the first of the two objections -- the a priori philosophical argument in favor of superficialism.

My second objection is more architectural and also more empirical. It's a guess or bet against a representationalist architecture according to which we literally store representational contents like "there's a beer in the fridge" and retrieve those contents when they are relevant to our reasoning.

Actually, there's a possibility of a partial reconciliation here, of a different sort. After all, what is it to literally have a stored representation with the content "there's a beer in the fridge"? Obviously, there's no literal slip of paper with that sentence written anywhere in the brain. Maybe a complex distributed process can count as literally storing that representation if certain other conditions are met.

Here, I think the representationalist faces a dilemma. On the one hand, the representationalist can be super liberal and say that there's a stored representation that "there's a beer in the fridge" whenever the system is such that it has the dispositional structure characteristic of a beer-in-the-fridge believer. In that case, there's no substantive empirical dispute between dispositionalism and representationalism: The so-called representationalist is really just a dispositionalist employing misleading language. On the other hand, the representationalist can make specific architectural commitments regarding how the cognitive system must be designed. The more specific, the riskier empirically, of course, and the closer to the simplistic cartoon sketch, and the less relevant to our superficialist interests as belief ascribers.

As participants in the Princeton workshop noticed, and as readers of the debate between Jake Quilty-Dunn, Eric Mandelbaum, and me (e.g., here) sometimes notice, although there's a bald top-line disagreement between dispositionalism and representationalism, a closer look suggests some paths toward reconciling the two views. However, I think a still closer look suggests that such apparent reconciliations can really only be partial. Understanding this back-and-forth helps us better understand the philosophical terrain and the real nature of the dispute.

The Continental/Analytic Divide Is Alive and Well in Philosophy: A Quantitative Analysis

A major sociological divide in recent Anglophone philosophy is the divide between philosophers who see themselves as working in the tradition of Nietzsche, Heidegger, Foucault, and Derrida -- so-called "Continental" philosophers -- and those working in the tradition of Frege, Russell, Wittgenstein, and Quine -- so-called "analytic philosophers". This division is reflected, in part, in journal citation patterns. You might wonder about this history of this. Was Philosophical Review always allergic to Nietzsche? Or is that a relatively recent phenomenon? And how extreme is the phenomenon? Do leading Continental figures get at least some play in the top analytic journals, or are they almost entirely excluded?

I first looked at this issue quantitatively ten years ago. Today's post is a reanalysis, with new and updated data.

The "big three" Anglophone philosophy journals -- all of which have been leading journals since the first decade of the 20th century -- are Philosophical Review, Mind, and Journal of Philosophy (formerly Journal of Philosophy, Psychology, and Scientific Methods). All currently lean heavily "analytic". Recent journal rankings also tend to classify Nous (founded 1967) as similarly prestigious. All are also indexed in JStor, along with a diverse group of 135 other philosophy journals, many of which are not as sociologically aligned with the analytic tradition.

What I've done is to look, decade by decade, at the rates at which the names of central analytic and Continental philosophers appear in these "big 4" journals compared to other journals.

Compare, first, Nietzsche and Frege -- foundational figures for both traditions, both born in the 1840s. The crucial measure is the percentage of articles in which each figure is mentioned in the big four vs. the percentage in which each is mentioned in the remainder. For methodological details see this note.[1] For a clearer view of the charts below, click them.

As you can see, through the 1940s or 1950s the Nietzsche lines stay more or less together and the Frege lines stay more or less together. This means that Nietzsche and Frege are mentioned about equally frequently in the big four (actually big three, back then) journals as in other journals -- Nietzsche in about 5% of articles throughout the period and Frege in about 2%.

Starting in the 1960s or 1970s, the dashed and solid lines are clearly separated -- a separation that increases through the 1990s. Since the 1990s, the separation has remained quite large. Note how often Frege is mentioned in the big four journals in the 1980s through 2010s -- in about 20% of all articles. Outside the big four, he's mentioned in about 6% of all articles. Also outside the big four, Nietzsche is mentioned about as often as Frege -- in about 5% of articles. But in the big four, he is mentioned in only about 1% of articles by the 2000s and 2010s.

In the period from 2000 to 2019, Nietzsche is mentioned only 20 times among 1760 articles in the big four. If you were to read every article published in the big four journals, you would see his name mentioned on average in only one article per year. That's remarkably infrequent for such a historically important figure!

A similar story holds for Heidegger and Wittgenstein -- leading early figures in the Continental and analytic traditions, respectively -- and both born in 1889. (Again, click chart for clarity.)

Starting in the 1950s, Wittgenstein is notably more favored in the big four than in the others, though the difference isn't extreme. Starting the 1940s, Heidegger is slightly disfavored by the big four relative to the others, with the difference getting large by the 1980s and continuing to increase up to the present. In the 2010s, Heidegger is mentioned in 0.7% of articles in the big four (5 times total in 764 articles) and in 6% of articles in the remaining journals (1474/26084).

Okay, how about the Continentals Sartre (b. 1905), Foucault (b. 1926), and Derrida (b. 1930) vs. the analytics Quine (b. 1908), Chisholm (b. 1916), and Putnam (b. 1926)? The graph is a little crowded but the following should be evident: The muted-color analytics show higher in the Big Three (solid lines) than in the remaining journals (dashed lines), while the bright-color Continentals show the reverse pattern -- and the spread is much more evident in the past few decades than it was mid-century. (There's a bit of false-positive noise for Foucault and Putnam, but not enough to mask the general trend. Russell I have chosen to exclude entirely due to false positives.)

Here's one way of thinking about it. In the bulk of JStor philosophy journals, representing a mix of journals with Continental vs. analytic vs. eclectic perspectives, these six figures are all broadly similar in importance in recent decades, as shown by the fact that the dashed lines end up all bunched in the middle range of about 2-7% in the most recent decades. But if we look at the big 4 analytic journals, the analytic figures loom large, especially Quine, while the Continentals are near the floor -- Derrida and Foucault in particular being almost 0% (each has 4 mentions total in the period from 2000-2019, i.e., about once in each journal across the whole period).

The effect is even clearer if we take averages of trend lines for the five Continentals versus the five analytics:

In the 2000s and the portion of the 2010s that has so far been indexed in JStor, the words Nietzsche, Heidegger, Sartre, Foucault, and Derrida appeared in 48 articles total among 1760 big four journal articles (2.7%). Thus, the big four journals have included, on average, less than one article per journal per year that even passingly mentions any of these five authors.

Maybe an analysis of leading Continental journals would reveal a similar trend, with Nietzsche and Heidegger mentioned in a substantial percentage of articles and Frege and Wittgenstein hardly mentioned at all -- or maybe not. But even if not, the exclusion of leading Continental figures from the top analytic journals shows that the Continental/analytic divide remains sociologically important.

ETA, 12:05 p.m.

Could it be a result of the fact that the big four journals don't publish much ethics and some of the Continental figures mainly had their impact in ethics? I don't think so. If we look for ethics-specific journals among the top journals of philosophy "without regard to area" in Leiter's most recent poll (a bit dated), only Ethics and Philosophy and Public Affairs rank among the top 25. A disjunctive search for Nietzsche or Heidegger or Sartre or Foucault or Derrida shows only 38 hits total for the 17 year period from 2000-2016 -- again about one mention of any of these five authors per journal per year.

------------------------------------------

[1] Names include a truncation symbol, e.g., "Nietzsche*", which includes "Nietzsche's" and "Nietzschean", except for "Putnam", to exclude false alarms for the publisher "Putnam's sons", and except for the disjunctive search described in the penultimate paragraph. Percentages are divided by a representative universe of articles including the search term "the". Only English-language articles are included. Reviews and minor documents are excluded by limiting the results to "articles" in JStor. Although the search terms run through 2019, JStor only covers through the mid-2010s for many journals, including the big four.

Infinite Puppetry

About a year ago, I argued in a blog post that "Everything You Do Causes Almost Everything". If the universe is temporally infinite (as suggested by the current default theories in cosmology) and supports random fluctuations post-heat-death (as also suggested by the current default theories in cosmology), then every action you take will perturb some particles, which will perturb more particles, which will perturb more particles, in an infinite causal chain (a "ripple"), eventually perturbing some post-heat-death system in a way that results in any type of non-unique, non-zero-probability event that you care to specify. Wave your hand, and you will trigger a ripple of perturbances through the cosmos that will eventually cause a far-distant-future duplicate of Shakespeare to decide that Hamlet needs a happy ending.

Philosopher of physics Jacob Barandes and I have collaborated on a draft chapter developing the idea in more detail, for my forthcoming book with Princeton, The Weirdness of the World. I think you'll agree that the idea fits nicely with the book title!

The draft chapter more carefully articulates the physical assumptions required for the almost-everything-causes-almost-everything idea to work -- and then it adds some new thoughts, specifically the infinite puppetry idea.

Below I share the three final sections of the draft chapter, on infinite puppetry. (For more on almost-everything-causes-almost-everything see last year's blog post or the full chapter draft.) Thoughts and comments welcome as always!

Signaling Across the Vastness

The following will also almost certainly occur, given our assumptions so far: On some far distant post-heat-death counterpart of Earth will exist a counterpart of you – let’s call that person you-prime – with the following properties: You-prime will think “right hand” after the ripple from the act of your raising your right hand arrives at their world, and you-prime will not have thought “right hand” had that ripple not arrived at their world. Maybe the ripple initiates a process that affects the weather which causes a slightly different growing season for grapes, which causes small nutritional differences in you-prime’s diet, which causes one set of neurons to fire rather than another at some particular moment when you-prime happens to be thinking about their hands. Likewise, there’s a future you-prime who would have thought “A” if you, here on our Earth, had held up a sheet with that letter and not otherwise. Indeed, infinitely many future counterparts of you have that property. You can specify the message as precisely as you wish, within the bounds of what a counterpart of you could possibly think. Some you-prime will think, “Whoa! Infinite causation!” as a result of your having raised your hand and would not have done so otherwise.

These message recipients will mostly not believe that they have been signaled to. However, we can dispel their disbelief by choosing the fraction who, for whatever reason, are such that they believe they are receiving a signal if and only if they do in fact receive a signal. We can stipulate that we’re interested in you-primes who share the property that when your signal arrives they think not only the content of the signal but also “Ah, finally that signal I’ve been waiting for from my earlier counterpart.”[1]

There’s a question of whether one of your future counterparts could rationally think such a thought. But maybe they could, if they had the right network of surrounding beliefs, and if those beliefs were themselves reasonably arrived at. We’ll consider one such set of beliefs in the final section of this chapter.

Infinite Puppetry

You needn’t limit yourself to ordinary communicative signals. You can also control your future counterparts’ actions. Consider future counterparts with the following property: They will raise their right hand if you raise your right hand, and they will not raise their right hand if you do not. Exactly which counterparts have this feature will depend on exactly when you raise your hand and how, since that will affect which particles follow which trajectories when they are disturbed by your hand. But no matter. Whenever and however you raise your hand, such future counterparts exist.

Your counterparts’ actions can be arbitrarily complex. There is a future you-prime who will, if you raise your hand, write an essay word-for-word identical with the chapter you are now reading and who will otherwise write nothing at all. Maybe that you-prime is considering whether to write some fanciful philosophy of cosmology, as their last hurrah in a failing career as a philosopher. They’re leaning against. However, the arriving particle triggers a series of events that causes an internet outage that prevents them from pursuing an alternate plan, so they do write the essay after all. (A much greater proportion[2] of such future counterparts, of course, will write very different essays from this one, but we can focus on the tiny fraction of them who create word-for-word duplicates of this essay.)

Let’s call someone a puppet if they perform action A as a consequence of your having performed an action (such as raising your hand) with the intention of eventually causing a future person to perform action A. (Admittedly, you might need to agree with the assumptions of this chapter to be able to successfully form such an intention.) You can now wave your hand around with any of a variety of intentions for your future counterparts’ actions, and an infinite number of these future counterparts will act accordingly – puppets, in the just-defined sense.

We recommend that you intend for good things to happen. This might seem silly, since if the assumptions of this chapter are correct, almost every type of finitely probable, non-unique future event occurs, regardless of your benevolent or malevolent intent right now. Still, there is a type of good event that can occur as a result of your good intentions, which could not otherwise occur. That’s the event of a good thing happening in the far distant future as a consequence of your raising your hand with the intention of causing that future good event. So let’s choose benevolence, letting good future events be intentionally caused while bad future events are merely foreseen side effects.

A deeper kind of puppet mastery would involve influencing a person’s actions through a sequence of moves over time and with some robustness to variations in the details of execution. This might not be possible on the current set of assumptions. Raising your right hand, you can trigger arbitrarily long sequences of actions in some future you-prime. But if you then raise your left hand, there’s no guarantee that a ripple of particles from your left hand will also hit the same you-prime. Maybe all the ripples from your right hand head off toward regions A, B, and C of the future universe and all the ripples from your left hand head off toward regions D, E, and F. Similarly, if you raise your right hand like this, the ripples might head toward regions A, B, and C, and if you raise it instead like that, they head toward regions G, H, and I. So there might be no future counterparts of you who do what you intend if you raise your right hand now and then do what you intend when you raise your left hand later; and there might be no future counterparts who will do what you intend if you raise your right hand now, insensitively to the particular manner in which you raise it. In this way, there might be no sequencing and no implementational robustness to your puppetry.

Sequential and robust puppetry might only be reliably possible if we change one of the assumptions in this chapter. Suppose that although the universe endures infinitely in time, spatially it repeats – that is, it has a closed topology in the sense we described in Section 1 – so that any particle that travels far enough in one direction eventually returns to the spatial region from which it originated, as if traveling on the surface of a sphere. Suppose, further, that in this finite space, every ripple eventually intersects every other ripple infinitely often. Over the course of infinite time each ripple eventually traverses the whole of space infinitely many times; none get permanently stuck in regions or rhythms that prevent them from all repeatedly meeting each other. (If a few do get stuck, we can deal with them using the n^m strategy of Section 4. Also the rate of ripple stoppage would presumably increase with so much intersection, but hopefully again in a way that’s manageable with the n^m strategy.) When you raise your right hand, the ripples initially head toward regions A, B, and C; when you raise your left hand, they initially head toward regions D, E, and F; but eventually those ripples meet.

With these changed assumptions, we can now find future counterparts who raise their right hands as a result of your raising your right hand and who then afterward raise their left hand as a result of your afterward raising your left hand. We simply look at the infinite series of systems that are perturbed by both ripples. Eventually some will contain counterparts of you who raise their right hands, then their left, as a result of that joint perturbation. In a similar way, we can find implementationally robust puppets: counterparts living in systems that are perturbed by your actual raising of your right hand (via the ripple that initially traversed regions A, B, and C) and which are also such that they would have been perturbed had you, counterfactually, raised your hand in a somewhat different way (via the ripple that would have initially traversed regions G, H, and I). Multiplying the minuscule-but-finite upon the miniscule-but-finite, we can now find puppets whose behavioral matching to yours is long and implementationally robust, within reasonable error tolerances.

We Might All Be Puppets

So far, we have not assumed that anything existed before the Big Bang. But if the universe is infinite in duration, with infinitely many future sibling galaxies, it would be in a sense surprising if the Big Bang were the beginning. It would be surprising because it would make us amazingly special, in violation of the Copernican Principle of cosmology, which holds that our position in the cosmos is not special or unusual. We would be special in being so close to the beginning of the infinite cosmos. Within the first 14 billion years, out of infinity! It’s as though you had a lotto jar with infinitely many balls numbered 1, 2, 3… and you somehow managed to pull out a ball with the low, low number of 14 billion. If you don’t like a strictly infinite lotto, consider instead a Vast one. The odds of pulling a number as low as 14 billion in a fair lottery from one to a Vastness are far less than one in a googolplex.[3]

Cosmologists don’t ordinarily deny that there might have been something before the Big Bang. Plenty of theories posit that the Big Bang originated from something prior, though there’s no consensus on these theories.[4] If we assume that somehow the Big Bang was brought into existence by a prior process, and that process in turn had something prior to it, and so on, then the Copernican lottery problem disappears. We’re in the middle of a series, not at the beginning of one. Maybe Big Bangs can be seeded in one way or another. Heck, maybe the whole observable universe is a simulation nested in a whole different spatial reality (Chapters 4 and 5) or is itself a very large fluctuation from a prior heat-death state.

Suppose, then, that we are in the middle of an infinite series rather than at the beginning of one, the consqeuence of accepting both Copernican mediocrity and an infinite future. If so, and if we can trace chains of causation or contingency infinitely backward up the line, and if a few other assumptions hold, then eventually we ought to find our puppeteers – entities who act with the intention of causing people to do what we are now doing and whose intentions are effective in the sense that had they not performed those actions, we would not be here doing those things. Suppose you are knitting your brow right now. Somewhere in the infinite past, there is a near-duplicate counterpart of you with the following properties: They are knitting their brow. They are doing so with the intention of initiating ripples that cause later counterparts of them to knit their brows. And you are just such a later counterpart, because among the events that led up to your knitting your brow, absent which you wouldn’t have knit your brow, was a ripple from that past counterpart.

We the authors of this chapter – Eric and Jacob – can work ourselves into the mood of finding this probable. An infinite cosmos is simpler, more elegant, and more consistent with standard cosmological theory; if it’s infinite, it’s probably infinite in all directions; and if it’s truly infinite in all directions, there will be bizarre consequences of that infinitude. Puppetry is one such consequence. We would not be so special as to be only puppeteers and never puppet. It seems only fair to our future puppets to acknowledge this.

--------------------------------------

[1] Compare this procedure with Sinhababu’s 2008 procedure for writing love letters between possible worlds. One advantage of our method over Sinhababu’s is that there actually is a causal connection.

[2] Here and throughout we bracket quibbles about ratios of infinitude by considering the limit of the ratio of counterparts with property A to counterparts with property B as the region of spacetime defined by your forward lightcone goes to infinity.

[3] Our reasoning here resembles the reasoning in the “Doomsday argument”, e.g., Gott 1993, according to which it’s highly unlikely that we’re very near the beginning of a huge run of cosmological observers. For a bit more detail, see Schwitzgebel 2022b. For another related perspective, see Huemer 2021.

[4] See notes 12 and 13 (in the full draft) for references. A note on terminology: “Prior” sounds kind of like “earlier” but is also more general in that there’s a sense in which one thing can be ontologically prior to another, or ground it, or give rise to it, even if they one doesn’t temporally precede the other (e.g., an object is prior to its features, or noumena are prior to phenomena [see Chapter 5 of the book draft]). Possibly, temporal priority is a relationship that only holds among events within our post-Big Bang universe while whatever gave rise to the Big Bang stands in some broader priority relationship to us.

[image source]

After a Taste-Bud Hiatus, Experiencing Candy Like a Six-Year-Old

I used to blog quite a bit about weird aspects of sensory experience, back when my central research interest concerned the vagaries of introspection and the strange things people say about their streams of experience. (See a few sample posts; some articles; two books.) I thought I'd share another today -- something striking to me -- though actually not that weird, I suppose.

About a month ago, I accidentally bit down hard on an unpopped popcorn kernel, "bruising" the teeth on the left side of my mouth. (Yes, that's a thing. My dentist tells me nothing is broken or cracked; it just needs time to heal.) It was remarkably painful to chew on that side, and for weeks I chewed entirely on the right side of my mouth, barely even letting food drift to the left side. Last week, I resumed gently chewing on the left again -- just soft things, carefully, experimentally. Having dessert one night, I was suddenly struck by how much sweeter the dessert tasted on the left side than on the right side. Remarkably sweeter. Different enough that the fact really jumped out at me, though I wasn't at all expecting or looking for it.

I was eating an "orange slice" candy. You know, one of these guys:

On the right side of my mouth, the candy was blandly sweet with a simple citrus flavor. On the left side, I experienced the candy as vividly sweet, zinging with orange. The contrast persisted as I moved the mass of candy around in my mouth. When I shifted the bulk to the right, it seemed to instantly lose flavor, like a piece of gum chewed too long. When I shifted it back to the left, the flavor brightened again.

I experimented with other candies over the next few days: lemon and lime slices, chocolate, peppermint sticks. I consistently found the left side sweeter than the right -- and not only sweeter, but also more vividly flavored in other ways. However, I found no similarly noticeable difference for savory flavors, or tea, or pure salt, straight lemon juice, or many of the other things I have eaten since. The effect was mostly or entirely limited to sweetness only, and the associated flavors of the sweet things.

I remember loving fruit slice candies when I was six. I would savor them for fifteen minutes, driving my parents nuts, who had to wait for me at the end of meals. (Now I tend to wolf down desserts: See my defense of dessert-wolfing.) The flavor of the orange slice resonated with my memories of youth. It was like my taste buds -- or the related sensory regions in my brain -- were six years old again. It seemed to me that the orange slice tasted to me now, on the left side of my mouth, in that amazing way it had tasted to me as a child, and then when I shifted it to the right side, it fell back into the blandness that I have since become accustomed to.

I'm not sure why the effect was limited to sweetness. In general, taste sensitivity declines with age, but the decline seems to be as strong for salty, savory, and bitter tastes as for sweet ones. My taste experience has probably dulled in multiple respects. Why sweetness only should rejuvenate, I have no idea -- even more confusingly, not simple sweetness only but the more complex flavors tangled up with sweetness, such as chocolate and sweet orange.

A week later, I find that the effect is still present, though diminishing. I want the vivid sweetness back! The experience acutely reminds me of how much of what we vividly experience recedes into a fog with ageing. A comparison point: I remember getting glasses as a teenager after years of slightly blurry vision and loving how sharp the world became. Now even the best prescription I can find will never make the world that sharp. I also feel that when I read fiction I don't quite as vividly imagine the scenes as I used to.

Middle age has its compensating advantages. I'm mellower, more settled. Even sensorily, things are open to me that weren't before: Presumably because of my diminished taste sensitivity, I can enjoy bitter coffee and sharp cheese. But the hiatus from left-side chewing, followed by some fleeting new candy raptures, has given a sharp new tang to my thoughts about sensory loss with age.

[image source]

Our Infinite Predecessors: Flipping the Doomsday Argument on Its Head

The Doomsday Argument purports to show, probabilistically, that humanity will not endure for much longer: Likely, at least 5% of the humans who will ever live have already lived. If 60 billion have lived so far, then probably no more than 1.2 trillion humans will live, ever. (This gives us a maximum of about eight more millennia at the current birth rate of 140 million per year.) According to this argument, the odds that humanity colonizes the galaxy with many trillions of inhabitants are vanishingly small.

Why think we are doomed? The core idea, as developed by Brandon Carter (see p. 143), John Leslie, Richard Gott, and Nick Bostrom is this. It would be statistically surprising if we -- you and I and our currently living friends and relatives -- were very nearly the first human beings ever to live. Therefore, it's unlikely that we are in fact very nearly the first human beings ever to live. But if humanity continues on for many thousands of years, with many trillions of future humans, then we would in fact be very nearly the first human beings ever to live. Thus, we can infer, with high probability, that humanity is doomed before too much longer.

Consider two hypotheses: On one hypothesis, call it Endurance, humanity survives for many millions more years, and many, many trillions of people live and die. On the other, call it Doom, humanity survives for only another a few more centuries or millennia. On Endurance, we find ourselves in a surprising and unusual position in the cosmos -- very near the beginning of a very long run! This, arguably, would be as strange and un-Copernican as finding ourselves in some highly unusual spatial position, such as very near the center of the cosmos. The longer the run, the more surprisingly unusual our position. In contrast, Doom suggests that we are in a rather ordinary temporal position, roughly the middle of the pack. Thus, the reasoning goes, unless there's some independent reason to think Endurance to be much more plausible than Doom, we ought to conclude that Doom is likely.

Let me clarify by showing how Doomsday-style reasoning would work in a few more intuitive cases. But first, here's an inverted mushroom cloud to symbolize that I'll soon be flipping the argument over.

Imagine two lotteries. One has ten numbers, the other a hundred numbers. You don't know which one you've entered into, but you go ahead and draw a number. You discover that you have ticket #6. Upon finding this out, you ought to guess that you probably drew from the ten number lottery rather than the hundred number lottery, since #6 would be a surprisingly low draw in a hundred-number lottery. Not impossible, of course, just relatively unlikely. If your prior credence was split 50-50 between the two lotteries, you can use Bayesian inference to derive a posterior credence of about 91% that you are in the ten-number lottery, given that you see a number among the top ten. (Of course, if you have other evidence that makes it very likely that you were in the hundred-number lottery, then you can reasonably retain that belief even after drawing a relatively low number.)

Alternatively, imagine that you're one of a hundred people who have been blindfolded and imprisoned. You know that 90% of the prison cells are on the west side of town and 10% are on the east side. Your blindfold is removed, but you don't see anything that reveals which side of town you're on. Nonetheless, you ought to think it's likely you're on the west side of town.

Or imagine that you know that 10,000 people, including you, have been assigned in some order to view a newly discovered painting by Picasso, but you don't know in what order people actually viewed the painting. Exiting the museum, you should think it unlikely that you were either among the very first or very last.

The reasoning of the Doomsday argument is intended to be analogous: If you don't know where you're temporally located in the run of humans, you ought to assume it's unlikely that you're in the unusual position of being among the first 5% (or 1% or, amazingly, .001%).

Now various disputes and seeming paradoxes arise with respect to such probabilistic approaches to "self-location" (e.g., Sleeping Beauty), and a variety of objections have been raised to Doomsday Argument reasoning in particular (Leslie's book has a good discussion; see also here and here). But let's bracket those objections. Grant that the reasoning is sensible. Today I want to add a pair of observations that have the potential to flip the Doomsday Argument on its head, even if we accept the general style of reasoning.

Observation 1: The argument assumes that only about 60 billion humans have existed so far, rather than vastly many more. Of course this seems plausible, but as we will see there might be reason to reject it.

Observation 2: Standard physical theory appears to suggest that the universe will endure infinitely long, giving rise to infinitely many future people like us.

There isn't room here to get into depth on Observation 2. I am collaborating with a physicist on this issue now; draft hopefully available soon. But the main idea is this. There's no particular reason to think that the universe has a future temporal edge, i.e., that it will entirely cease. Instead, standard physical theory suggests that it will enter permanent "heat death", a state of thin, high-entropy chaos. However, there will from time to time be low-probability events in which people, or even much larger systems, spontaneously congeal from the chaos, by freak quantum or thermodynamical chance. There's no known cap on the size of such spontaneous fluctuations, which could even include whole galaxies full of evolving species, eventually containing all non-zero-probability life forms. (See the literature on Boltzmann brains.) Perhaps there will even be new cosmic inflations, for example, caused by black holes or spontaneous fluctuations. Vanilla cosmology thus appears to imply an infinite future containing infinitely many people like us, to any arbitrarily specified degree of similarity, perhaps in very large chance fluctuations or perhaps in newly nucleated "pocket universes".

Now if we accept this, then by reasoning similar to that of the Doomsday Argument, we ought to be very surprised to find ourselves among the first 60 billion people like us, or living in the first 14 billion years of an infinitely existing cosmos. We'd be among the first 60 billion out of infinity. A tiny chance indeed! On Doomsday-style reasoning, it would be much more reasonable, if we think the future is infinite, to think that the past must be infinite too. Something existed before the Big Bang, and that something contained observers like us. That would make us appropriately mediocre. Then, in accordance with the Copernican Principle, we'd be in an ordinary location in the cosmos, rather than the very special location of being within 14 billion years of the beginning of an infinite duration.

The situation can be expressed as follows. Doomsday reasoning implies the following conditional statement:

Conditional Doom: If only 60 billion humans, or alternatively human-like creatures, have existed so far, then it's unlikely that many trillions more will exist in the future.

If we take as a given that only 60 billion have existed so far, we can apply modus ponens (concluding Q from P and if P then Q) and conclude Doom.

But alternatively, if we take as a given that (at least) many trillions will exist in the future, we can apply modus tollens (concluding not-P from not-Q and if P then Q) and conclude that many more than 60 billion have already existed.

The modus ponens version is perhaps more plausible if we think in terms of our species, considered as a local group of genetically related animals on Earth. But if we think in terms of humanlike creatures instead specifically of our local species, and if we accept an infinite future likely containing many humanlike creatures, then the modus tollens version becomes more plausible, and we can conclude a long past as well as a long future, full of humanlike creatures extending infinitely forward and back.

Call this the Infinite Predecessors argument. From infinite successors and Doomsday-style self-location reasoning, we can conclude infinite predecessors.

---------------------------------------------

Related:

Almost Everything You Do Causes Almost Everything (Mar 18, 2021)

My Boltzmann Continuants (Jun 6, 2013).

How Everything You Do Might Have Huge Cosmic Significance (Nov 29, 2016).

And Part 4 of A Theory of Jerks and Other Philosophical Misadventures.

[image adapted from here]