Intuition Pumps and Other Tools for Thinking

This book is misleadingly titled  – it should be called Intuition Pumps and Other Tools for Thinking like Daniel Dennett. But no harm done.

Having previously only heard of Daniel Dennett in context of his status as one of the four horsemen of the New Atheism movement, I picked up this book from Kinokuniya blind – expecting it to be a general sort of empirically-minded self-improvement book for honing your thinking techniques. Instead, I got a thinly-veiled overview of Dennett’s various philosophical positions over meaning, consciousness, complexity, and free will, which are explained using such thinking tools as Dennett recommends that the reader cultivate. The subtext being that if the reader were only to apply the correct thinking tools, she would naturally reach the same, or similar, conclusions, that Dennett has.

The central thinking tool in the box is really the intuition pump, a kind of thinking experiment cunningly worded and framed to get the reader to come to a desired conclusion by appealing to his or her intuitions. The term was coined by Dennett and many examples are prominently featured in this book, given such whimsically memorable names as the Curse of the Cauliflower and the Zombic Hunch. A famous example of an intuition pump (and one Dennett savages) is the Chinese Room experiment – in which one imagines a man using a set of algorithmic translation rules to translate Chinese questions into English in order to answer them. However, at no point in this process does the man actually understand the inputs – he merely acts upon them based on a set of instructions. The point of this thought experiment is to pump the intuition that – like the man in the Chinese room – computers that use algorithms to process data do not truly understand them – ergo, consciousness (i.e. understanding) cannot arise out of a set of algorithms designed to simulate its functions.

The intuition pump is really a double edged sword, however, and can be used to both illuminate and mislead. This is because intuition pumps are by their very nature abstractions of philosophical problems – they simplify conditions to make it easier to think about a thing – and in some cases, to think about a thing the way the creator intended. Dennett tells us to interrogate intuition pumps by turning the knobs – in other words, changing some of the conditions or assumptions of the intuition pump and seeing if the same intuitions are pumped out. Even as Dennett presents his own intuition pumps, he also deconstructs pumps that have been devised by others. He calls misleading intuition pumps boom crutches, deliberately malappriopriating (see what I did there?) the nautical term for these anti-thinking tools that are designed to shut down thinking.

Ultimately, however, after a few short chapters laying out his toolbox, Dennett begins the methodical quest of introducing the reader to his brand of philosophical thought. Dennett’s positions on various matters – cognition, free will, consciousness – can be boiled down to the idea that these ideas – held by some philosophers to reside in the sacred space of epistemological inscrutability- can actually be explained with a mundane scientific approach. The hard questions are procedurally hard – but not unsolvable using scientific inquiry. In plain language, it simply means that things like consciousness or free will can be explained or described in the framework of material reality, and there is no need to posit some imagined higher substrate of reality – such as positing that consciousness manifests beyond the physical mind – in order to understand such phenomena.

It’s hard to distill Dennett’s ideas into one paragraph, but an example of this approach is Dennett’s deconstruction of the Chinese Room. That thought experiment holds that consciousness cannot manifest out of algorithms, which are ostensibly what the mind, in Dennett’s model of cognition, uses – i.e. strong, conscious AI is not possible. Dennett’s dictum, on the other hand, is that software, sufficiently advanced (and operating on a sufficiently complex hardware stratum like the brain), approximates consciousness to such a level of fidelity that it is indistinguishable from consciousness, and in fact, is consciousness, and in actual fact is precisely what consciousness is in reality.

Think about the idea of the philosophical zombie – a being that acts and talks so like a human being that it is indistinguishable from one, except that it has no internal world, no consciousness – and elevate it to the idea of the zimboe – another philosophical creature that simulates second-order beliefs that it is a conscious thinking being, when it actually isn’t. A zimboe looks, acts, and feels – or thinks it feels – like a real human being – but it doesn’t actually think anything. Dennett’s argument is that the zimboe is the logical outcropping of the idea that there can be p-zombies that approximate human complexity, but at that level of fidelity it would be impossible for a zimboe to be distinguished from a human being, by any means. Rather than posit some ethereal essence of consciousness, Dennett argues, why not just eliminate the need for that as-yet-undiscovered substrate and go straight to the Occam’s-Razor solution – that the zimboe concept is incoherent because a sufficiently complex bundle of computational processes – i.e. the zimboe – already possesses, by dint of its mental complexity, consciousness?

Back to the Chinese Room – Dennett’s point is that the flaw of the intuition pump is that it doesn’t specify how complex the algorithms operated upon by the man must be to make it so that the questioners who are asking the questions in Chinese are satisfied that the answers they get are from a thinking entity. Dennett’s intuition is that they would have to be astoundingly complex – more complex than would be possible for one man and a pencil to analyse in any reasonable time horizon. As such, the Chinese Room thought experiment is bust because it fails to account for the possibility that the set of algorithmic instructions would only be physically possible to operationalise on a substrate capable enough to elevate it to a level in which it could be described to “understand” its instructions – which could approximate something like consciousness.

Dennett applies this same set of thinking tools to different fields – thinking about meaning, evolution, free will. In each case Dennett purports to banish the veil of mysticism that some drape over seemingly intractable problems, and shows that they are really resolvable with the application of tried-and-true thinking methods. Another interesting example is Dennett’s stance on free will. Dennett is really a compatibilist – he holds that free will is compatible with determinism. How? Simple – the kind of philosophically ideal free will that some imagine we must possess to rightly be called free is an incoherent concept. Barring quantum indeterminacy, at some level the rules of physics determine the course of the universe – that is determinism. However, reality is so complex that the kind of “free will” some would characterise as being required to be truly free is not required to make meaningful decisions. Dennett characterises the difference as being between a sufficiently advanced pseudo-random number generator (one that generates a random list of numbers using a series of step by step algorithms) and a “true” random number generator. Sufficiently advanced, there is no compelling reason to choose the latter, unless you plan to play chess with an omniscient God that can predict your every move. But who needs, or wants, that kind of ideal freedom? Free will is not a binary – it is a spectrum, and what free will we have is sufficient for our purposes down to the subatomic level.

I’ve gone off on a bit of a tangent here – but I should be somewhat embarrassed to admit that Dennett gives voice to some of my own intuitions about mind, consciousness, and free will. I find myself agreeing with him to a very large and almost embarrassing extent. Far from being accused of being overly impressionistic, however, I should caveat that the irony of my reading experience of Dennett’s tome is that I didn’t glean my philosophical insights from his intuition pumps, but really from how my own intuitions corroborated with his conclusions. Dennett, here, was in effect preaching to the choir. And even more ironic was that I came away with a sneaking suspicion that his intuition pumps were simply not especially effective in conveying the point in  a watertight manner. But I suppose that is a suspicion that fosters itself. Intuition pumps are by nature unwieldy thinking tools – ripe for abuse or misuse, abstracted in ways that cloud judgment, with too many hypotheticals and variables. Are Dennett’s own pumps immune to this kind of thing? Probably not. He has an easier job of it though – his positions are fundamentally more parsimonious – after all, he is arguing from the standpoint of “possibility” – which requires less epistemological certainty than to declare a thing “impossible”. It is possible to conceive of consciousness, evolution and free will in terms of physical laws, and this is how it is possible to conceive of them. They may be the wrong models – perhaps consciousness does exist as a floaty-substrate, by Dennett’s own admission. But Dennett doesn’t have to show that his model is correct – he just has to show that it is possible to think of them as being grounded in empirics. His critics often have to show that a thing is impossible to characterize in terms of physical reality – which requires some boom crutch-y sleight of hand and other such mental breakdancing. It is impossible that consciousness is rooted in the physical world – here is inductive proof in the form of a hypothetical! Forgive me if I don’t find that particularly convincing.

But I ramble. Dennett’s book can describe all these jumbled thoughts better than I can. Go read it, regardless of your metaphysical inclinations. I think you’ll probably get something out of it. At the very least, the chance to turn some of Dennett’s knobs at his invitation.

I give this book: 4.5 out of 5 Sphexes

Broken Age


Broken Age is like a jacuzzi – bubbly and fun but not very deep.

Broken Age is a classic point-and-click adventure game that also happens to be one of gaming’s most notable crowdfunding success stories. The second fact happens to overshadow the first. After all, game creator Tim Schafer and his team were responsible for some of the most beloved adventure game classics of the last century – Grim Fandango, Escape from Monkey Island, and Day of the Tentacle, to name a few – and the game amassed over $3.45 million in support – almost ten times as much as its initial goal of $400,000. On my part, I think Broken Age, while no instant classic, is a game that can stand on its merits, even as it is somewhat unfairly judged in relation to its antecedents and the stratospheric expectations that came with its crowdfunding success.

There is a pleasing symmetry in the game’s overall story. From the onset, the game tells the parallel stories of two protagonists, Vella and Shay: teenagers who have had overbearing social expectations placed upon them, and who struggle to escape those societal obligations. The player can switch between the stories at any time, but otherwise the parallel narratives do not intersect in the first half of the story. Where it gets more interesting is when the big reveal at the end of the first Act twists the plot strings into something resembling an hourglass, and the protagonists find themselves in each other’s worlds, so to speak, and their respective stories and puzzles begin to have an influence on each other. But this does mean that the story is relatively constrained – the second Act is a retread of the same locations and characters encountered in the first Act.

It’s a smart approach from a design perspective – the game’s puzzles start out relatively straightforward, but ramp up in complexity as the two narrative strands start to influence each other, requiring players to switch back and forth between protagonists to solve puzzles and propel the story forward. The puzzles themselves are actually mostly also quite well-designed, as far as point-and-click adventure games can be. In my review of Schafer’s earlier magnum opusGrim Fandango, I expressed some frustration about how the puzzles in Grim Fandango often crossed the threshold from challenging to frustratingly obtuse, trial-and-error mix-and-matching. I’m pleased to report that this is not a significant problem in Broken Age. Even the game’s most complex puzzles are logically laid out and rely more on thinking and memory than fruitless pixel-picking. Unlike Grim Fandango, the puzzles of Broken Age are actually fun to work out.

Less salutary, however, is the fact that the puzzles sometimes make no sense in the context of the game’s plot. In the second half of the game, the protagonists are made to solve puzzles that they would have no way of knowing the solution for, because the solution lies in clues only found in the other character’s game world. Of course, the player knows, since they play both characters, but in the context of the story, the puzzles rely on information that neither of them could possibly have had. This actually highlights that old gameplay vs narrative tension. In this case, the game’s puzzle designers clearly tried to privilege gameplay over plot, to require that players switch between worlds to solve puzzles. In all fairness, it would have been much harder to make it work from a narrative point of view, and given the ‘floaty logic’ that is par on the course for most adventure games, it can be forgiven in the light of time and budgetary constraints.

Speaking of floaty logic, the world is positively Schaferesque in how it is so delightfully free of internal consistency. Like Grim Fandango’s intriguing but nonsensical Land of the Dead, there is little sense that the world of Broken Age would make sense if suddenly ported over to reality. Anachronisms abound – Vella, who grew up in a village seemingly without any apparent technology, seems oddly au fait with the notion of video games and robots. Shay, who has grown up in a spaceship all his life, with no other human companionship, seems strangely cavalier about exploring a natural environment and interacting with others in the second half. There are many characters that seem out of place with their milieu – the hipster lumberjack in a forest of talking trees is one amusing example. Some plot elements are never really explained – like the fact that Shay didn’t find out his parents were real people instead of computers until the second act (I still don’t see how that could possibly make any sort of sense – and this is really the game’s biggest plot flaw). The world is static – there is no sense of time, history, place – it is a fairytale canvas upon which the plot unfolds like a children’s bedtime story. But given that the game really is just like a kid’s story book, it isn’t a deal-breaker. And the environments and music are beautifully rended in painterly 2D fashion, which complements that fairy-tale aesthetic. ‘Schaferesque’, then, is not really a pejorative here – merely a creative direction, and that’s fine.

This is because a Tim Schafer game is really made by its dialogue. If Tim Schafer knows how to do something, it’s the kind of all-ages humor endemic to adventure games, with just the right amount of glibness, satire, and heart without crossing over to being trite. Although his overall effort here doesn’t quite match up to certain parts of Grim Fandango, it’s still an inspired effort, especially compared to the dross of the rest of the video game industry. It’s actually worth playing the game just to listen to its dialogue, which happens to be delivered by some truly talented voice actors. My personal favorites are the snarky cloud-town denizen M’ggie and the talking spoon and knife (don’t ask) from Shay’s spaceship.

Ultimately, however, the interactions between characters, while fun, don’t elevate the plot or characters of the game to true classic status. While Grim Fandango had Manny Calavera, who was suave but conflicted and plagued with an uneasy conscience, who spent his years working towards physical and emotional salvation, the characters of Broken Age are still bit players in the two dimensional plot, marionettes compelled to follow the rigid parallel-hourglass structure of the game. In the first Act, the motivations of Shay and Vella are established – Vella wants to escape her fate as a maiden offered in sacrifice to a chthonic monster-deity, and Shay, smothered by the attentions of his overbearing and infantile “Mom” computer, wants to break out of the endless, sanitized routine of his daily existence. Those motivations aren’t really borne out in the second half – Shay and Vella do what they do because plot exigencies compel them to. Instead of being constrained by their circumstances, they become constrained by plot, carrying out fetch quests to set the stage for the final Act, at which point they actually get to be the heroes.

But even this final Act has its own problems. The game has an overarching and not especially subtle message – that people should connect and that good things come out of communication and talking to each other, to size up each other as people, not ciphers or things, and to live unconstrained by the obligations unfairly imposed upon you. However, in the denouement of the story, Shay and Vella have to work together to solve the final puzzle – and they do – but the ironic thing is that they never actually talk to each other – they just know what to do by virtue of the gameplay trick of the player knowing both parties’ circumstances. Shay and Vella never actually speak to each other throughout the entire game – and that’s where the thematic charge of Broken Age falls through into self-parody. They don’t actually communicate – they trust in their own guts to do what is right – but of course, they couldn’t have known what was right without talking in the first place. The game ends happily, in effect, without anyone talking to and working with each other – that social glue is purely an artifact of the player’s singular agency as the person who controls both characters. That’s the insuperable contradiction resulting from the gameplay-story tension of adventure games.

As such, Broken Age isn’t really more than the sum of its (well-made) parts. It’s a well-written and fun puzzler with a great aesthetic and high production values, but whose plot and themes don’t quite bear out in a way that provides the game with a thematic unity that goes with its design symmetry – in fact, it’s the opposite – its design and gameplay are often at odds with its themes and story. The dialogue bits are fun but they don’t add up to a particularly compelling narrative. The game doesn’t possess that ineffable element that would have made it a classic – but it’s hard to blame it for that. The odd thing about Broken Age is that people expected it to be a classic based on its pedigree and its kickstarter success, and are now savaging it because it doesn’t quite live up to those lofty expectations. It’s somewhat unfair, because, on its own merits, Broken Age is a great game that will give you a solid 10 hours of fun, even if it doesn’t astound you with its depth.

And those Hexapals are adorable.

I give this game: 3.5 out of 5 Superconducting Gyroscopic Hypercams