Picture of author.
3 Works 1,528 Membros 25 Críticas 1 Favorited

About the Author

Duncan J. Watts is associate professor of sociology at Columbia University and an external faculty member of the Santa Fe Institute. He holds a Ph.D. in theoretical and applied mechanics and has published in leading physics and sociology journals. He lives in New York City

Obras por Duncan J. Watts

Etiquetado

Conhecimento Comum

Membros

Críticas

Well written, about the ways that common-sense doesn't really apply to complex social situations - but why we cling to erroneous common-sense explanations anyway.
 
Assinalado
steve02476 | 11 outras críticas | Jan 3, 2023 |

The first parts of Everything is Obvious got the wheels in my brain spinning. The nuances AI is trying to understand how to use the NYC subway .... How the Mona Lisa became the greatest painting in the world William Shakespeare the best writer.

Each chapter started with so many possibilities. Especially as we moved into the later chapters, the narrative became more theoretical and less, well ... narrative. I was reading a lecture without the grounding of talking about people and their stories.

And at the conclusion, I don't really know what I read ... except well, now I have a better idea why the Mona Lisa is so famous.
… (mais)
 
Assinalado
wellington299 | 11 outras críticas | Feb 19, 2022 |
This is frequently described as a book on common sense, which it is, but more importantly it's an investigation on human cognitive limits more generally and also a call to radically restructure the discipline of sociology in light of modern advances in technology. Sociology often gets made fun of in the hierarchies of academic disciplines, but Watts argues that there are reasons why sociology seems so vague and unscientific: not only are sociological problems very complicated in ways that physics problems like orbital mechanics are not, but in addition to the fact that only now do we have the ability to run experiments to truly test our long-held prejudices about ourselves and how society works, our problem-solving skills are themselves subject to those same prejudices. It's a tall order, and though inevitably the chapters pointing out problems are stronger than the chapters suggesting ways to do better, I think this is an excellent synthesis of a lot of good information and a solid guide to outlining future research directions.

That human beings have cognitive biases is well-known, especially to readers of any Dan Ariely or Daniel Kahneman book, but the thing about them is that even if you know what they are and how they work, you're almost guaranteed to fall prey to them constantly anyway. "Common sense" is a powerful tool for navigating the complexities of life, but common sense is often just shorthand for a set of fallible mental shortcuts whose workings are almost invisible to us, and whose failures are only excusable by the fact that everyone else has all the same failures too. We use many heuristics to guide us through life, and those rules of thumb are often incoherent (Watts gives examples of proverbs that contradict each other like "look before you leap" vs "he who hesitates is lost"). This extends even down to the level of deeply held and supposedly universal beliefs about justice - when you play the ultimatum game or other simple exercises in game theory with people from different cultures, people behave in strikingly different ways due to cultural norms, and those cultural norms are themselves very difficult to clearly articulate or justify. His brief discussion of the extent to which what we think of as universal institutions like the market system vary dramatically throughout time reminded me a bit of Karl Polanyi's insights about how embedded capitalism is within culture and how unnatural in a way that is.

But the main issue is that "common sense" simply isn't designed to solve the kinds of sociological problems we now find ourselves encountering. Here Watts goes on a brief tour of some sociocultural phenomena like performance-based financial incentives, and how baffling the evidence is that they do anything at all. Not only are the effects relative to your peers (i.e. a $10,000 bonus can still be disappointing if everyone else got $20,000), but they're also relative to where you were before, and the bar for what prompts additional effort can keep being raised. Even high-value bonuses in both relative and absolute terms can have little effect on performance if what's being measured is unclear or easily gamed (think teachers being paid more for student achievement or Wall Street bankers paid for paper profits), and yet even after mounds of evidence undermining the case for simple performance metrics, it is guaranteed that you will hear someone think that the "common sense" insight that paying someone more will automatically result in better quality is essentially irrefutable.

Other examples are no less interesting. Watts poses some simple questions about the Mona Lisa: what makes the Mona Lisa the most famous painting in the world? Would the Mona Lisa's qualities have been apparent at 1700 when it was painted? What are the features that we currently consider it as having that no other painting does? Are there other paintings that share similar qualities, whether by da Vinci or someone else? Why aren't they as famous? Steady investigation shows that attempts to justify the painting's #1 ranking are either specious, examples of circular reasoning - "the Mona Lisa is the best because it has the qualities of the Mona Lisa [e.g. style, composition, brushwork, the smile, etc] and not something else" - or simply arbitrary, because as he shows, the Mona Lisa was not actually acclaimed with its current status as Best Painting Ever until a dramatic theft attempt in the early 20th century. The real reason it's the Primo Painting is basically that SOME painting has to be, and it really is arbitrary to some extent which one ends up with the top spot.

This has dramatic implications for any field where ranking depends on somewhat subjective factors (i.e. almost all fields). Music immediately leaps to mind, and Watts relates experiments he's run where people are asked to rank random songs both independently and also with the ability to see what other people have ranked those songs. Unsurprisingly, herd behavior and "trendiness" arises immediately when the equivalent of a Billboard chart is introduced to the experiment, which is something I've noticed myself when using software like last.fm that provides statistics on the music I listen to. Social network technology can just as easily be used to reinforce traditional hierarchies as to eliminate them. People ending up liking things simply because other people like them, and the implication is that many universally acclaimed bands are acclaimed not so much for any intrinsic merit as simple network effects. The same logic, with the slight complication of timing, extends to other "why this and not that" cases like Facebook's success and MySpace's failure, Minitel and the Internet, VHS and Betamex, etc etc. A slight initial random push might be enough to one product the edge over another in the cumulative advantage race, and only retrospectively are people able to offer countless competing and equally arbitrary theories on what that initial push was.

One implication of this line of reasoning that seems to disturb people is that a lot of life, including huge multi-billion dollar phenomena like why Harry Potter is so popular and not so many other superficially very similar YA series, is basically random. To put it another way, outcomes in a wide range of human endeavors that seem to depend greatly on human initiative follow simple statistical distributions that can also describe things like the outcomes of coin flips. What does that say about the common sense understanding of our own "specialness" or of our intuition that the world is divided into a few very influential people and many ordinary people? What does that say about our ability to predict the future to the extent that we follow strategies that leverage "specialness", as in trying to find "the next Harry Potter" or "the next Apple", or by trying to advertise to influential people in the hopes that they will influence their followers? After all, if you knew just the right social levers to push, you could do just about anything. Since we all know that special people are out there, waiting to be found, how do we identify their specialness and find them?

The problem is that in many cases, the special people, or the levers of history, are only able to be identified after the fact. As an example, Watts picks on Malcolm Gladwell for trying to figure out why Paul Revere is so famous while a guy named William Dawes, who went on a seemingly very similar ride at the same time, is virtually unknown today. Gladwell says that Revere was a "connector", a man unusually well-suited to his task of warning all the people on his route as opposed to the undistinguished Dawes; Watts says that nothing about Revere's current fame was destined at all, and if their routes had been swapped there's absolutely no reason to think that we wouldn't have identical "one if by land, two if by sea" poems about Dawes instead. Revere was simply in the right place at the right time, and it was only after the fact that people decided there must have been something unusual about him and his place in such a dramatic event in US history. The same story holds true with music or books: every book publisher in the world would love to be able to find "the next JK Rowling", but all save the lucky one couldn't even find the original JK Rowling, who was rejected many times. To many ostensibly well-trained people, there just wasn't anything about her work that seemed to stand out among the countless manuscripts of fantasy young adult novels they read every year (related quick prediction: Rowling's new non-fiction novel is probably pretty decent, but has a 0% chance of ever being placed in the canon alongside her Harry Potter work).

This is at root due to the fact that our brain is hard-wired to look for patterns and narratives even in realms where those metaphors are fundamentally inapplicable. History is another great example. Take the idea of the storming of the Bastille being a central event in the French Revolution, worthy of becoming the central national holiday of France. Could someone have known at the time that that particular event among all the chaos of the Revolution would have been so influential? Obviously not, but this means that all history is essentially a competition in storytelling, and that prediction in the Laplacian sense of perfect foreknowledge is impossible even if Newtonian physics were true. This notion has obvious relevance for important institutions like futures markets or business more generally. Rather than repeat myself, I'll just say that in the business sections Watts reiterates that people fall prey to all the same issues of mis-narrating history, learning the wrong lessons from the past, engaging in circular reasoning about why certain things are successful, believing that people are more special than they are, and assuming due to the halo effect that what is doing well now must have all sorts of other great attributes. (As a related note, David Romer had an excellent paper in 2006 called "Do Firms Maximize? Evidence from Professional Football" discussing how coaches systematically fail to maximize their expected points by failing to go for it on 4th down, simply because there are prevailing irrational norms about what constitutes acceptable risk; this predates the infamous Patriots 4th-and-2 against the Colts in 2008 but is still worth reading).

I'm not sure that all types of prediction are necessarily equal; off the top of my head, I would say that a book like The Limits to Growth, with its carefully-sourced numbers and logical formulae, should be looked at as a more credible forecast than a hedge fund prospectus. However, even if most attempts at predicting the future fail, and it seems best to just stick to simple models like always betting on the home team (which at a 58% success rate is within 3% of the best and most sophisticated algorithms you can concoct), there are important things you can do to reduce the failures, and hence resolve some of the issues raised in the rest of the book. It's not like you can act like the future is completely random; we have a drive to speculate for a reason. Here is where Watts is, predictably as it were, a little less helpful. Some useful sanity-check tools are aggregation of knowledge as opposed to relying on too few sources of information, encouraging experimentation rather than blindly staying the course, relying on local knowledge where possible rather on too much top-down direction (here Watts has a more level-headed take on this Hayekian principle than Tim Harford did in Adapt), and always trying to rely on measurement when rather than on intuition and "common sense" even if this is ultimately a somewhat Sisyphean goal.

With that, Watts transitions to the Big Picture. Knowing that common sense can mislead us is all when and good, but what new principles can we use to guide us in the future? Societies are big and complicated things, and merely saying that we can't trust common sense isn't good enough, especially when it comes to subjects like justice and fairness. So Watts moves in the direction of Justice as Fairness, explicitly advocating a Rawlsian view of designing institutions to maximize equity, contra Nozick. I agree with him, and I agree that the concepts explored in the book support the idea that social institutions should be designed with the least-well-off in mind, as well as that the "lemon socialism" behavior of the wealthy lately of acting like you are a special person on the upside and a helpless victim on the downside is both offensive and unjust. Acknowledging that we are members of a society and not a million Masters of the Universe is an old insight, but well-placed here, because now that we are beginning to have the technology to measure and analyze trends and social movements in great detail and in real time, we are also beginning to be able to subject foundational questions of justice to statistical analysis. Alexander Volokh had a great law review article in 1997 titled "n Guilty Men" which analyzed different societies' takes on the concept of "it is better to let X guilty people escape justice than to let one innocent person be punished" - Watts would argue that we are getting to the point where we could try to actually calculate that normative value.

Obviously that grandiose dream has many precedents - Watts mentions Auguste Comte, as one of many - and even more detractors. The idea that you could calculate something like justice seems absurd. Yet it certainly seems like more and more touchy subjects of the past are being re-examined with something approaching a scientific spirit. Dan Ariely and Michael Norton published a study in 2005 titled "Building a Better America-One Wealth Quintile at a Time" that polled people about income inequality that asked people what they thought a fair distribution should be. The results showed that most people, even rich people, supported a much more equal society than the one we have now. What role should that result play in discussions over redesigning the tax code? Is the only appropriate avenue for discussion about income during salary negotiations between each individual un-unionized employee and a hiring manager, or can/should we design broader institutions to better-implement more scientific notions of justice? These and many other questions that before belonged in the backs of philosophy and sociology books are finally able to be looked at with data, analysis, and experimentation - this book is a great overview of some good questions and hints of their answers. I'll be thinking about it for a while.
… (mais)
 
Assinalado
aaronarnold | 11 outras críticas | May 11, 2021 |
In de categorie Popular Science kan je vanalles tegenkomen, brol en kwaliteit. De flaptekst van 'Everything Is Obvious' sprak me wel aan, gezien m'n soms kritische en analytische kijk op bepaalde dingen, gebeurtenissen, enz.

In dit boek probeert Watts aan te tonen dat ons gezond verstand het niet altijd bij het rechte eind heeft, dat we vaak gebeurtenissen en dergelijke verkeerd inschatten. Of indien niet verkeerd, dan wel onvolledig. Een recent voorbeeld is de stakingen de afgelopen weken en waarom bijvoorbeeld (haven)arbeiders naar andermans gevoel en inschatting dom of marginaal overkomen. Dat kan liggen aan de karakters zelf, of hun omgeving, hun opvoeding, of wat dan ook. Maar voor de meerderheid van de massa (en voor de media ook, want dat levert kijk- en verkoopcijfers op) worden ze d'office als "minderwaardig" aanzien. Idem voor andere bevolkingsgroepen of soorten beroepen.

Watts schrijft ook hoe gezond verstand niet altijd leidt tot goeie oplossingen voor bedrijven die willen investeren in nieuwe producten of diensten en zich eventueel een voorspelling maken van de toekomst. Hoe het op zich wel goede denkpatronen zijn, maar dat men zich de kaas van het brood laat eten door een onverwachte wending van de markt, bijvoorbeeld. Voorbeelden zijn Sony (Betamax vs VHS, Minidisc vs Internet, enz...). En dan komt de vraag: als het wel goed gaat, ligt het aan de CEO? Aan de samenwerking in het bedrijf? Aan de markt zelf? Waaraan?

Over bijv. CEO's gesproken: Waarom worden quasi steeds dezelfde mensen gekozen voor een topjob? Of dezelfde soort mensen? Wat met degenen die evenzeer capabel zijn, maar door omstandigheden niet dezelfde kansen krijgen?

Ook bekendheid is een element dat aan bod komt: succes vs talent. Het is niet omdat al die beroemdheden in de muziek, film, sport, enz... zo succesvol zijn, dat ze ook getalenteerd zijn. Want er zijn ook veel getalenteerden die nooit het succes hebben van de anderen. Daarmee zijn vaak andere factoren gemoeid, zoals de consument/fan/supporter/... die zijn/haar voorkeur voor product x of persoonlijkheid y uitdraagt naar anderen toe en zo enige invloed bewerkstelligt.

En zo zijn er nog een aantal zaken die Watts in z'n betoog verwerkt. In ieder geval, zeker de moeite waard om te lezen.

Later in het boek gaat het eerder over sociologie zelf, hoe dat op eenzelfde manier als fysica, biologie, wiskunde, astronomie, enz. probeert wetten te vinden in haar onderzoek naar het gedrag van de mens, maar hoe dat tezelfdertijd wellicht niet zal lukken, omdat de wereld of het gedrag van de mens nu eenmaal geen eenvoudig of rechtlijnig gegeven is. Dat Watts daarvoor ook rekent op (bijv.) Facebook, is wellicht niet de beste manier.

Tot slot: Wordt Watts betaald door Amazon per keer hij het over Mechanical Turk heeft?

Het boek leest best wel vlot, ook al is het geen eenvoudige materie. Het is zeker interessant voor wie z'n kritische geest wat wil aanscherpen en bijwerken, om even een beter beeld op de maatschappij en menselijk gedrag te hebben; ook voor zichzelf, al naargelang. Wellicht bestaan er betere boeken (als ik zo bepaalde andere recensies lees), maar dat "probleem" heb je overal.

Als niet-specialist vind ik 'Everything Is Obvious' een eye-opener, ook al was niet alles "nieuw" voor me; ik dacht er alleen niet al te vaak bewust over na, over die bepaalde aangehaalde voorbeelden. Dus 'how common sense fails' is best een passende ondertitel. ;-)
… (mais)
 
Assinalado
TechThing | 11 outras críticas | Jan 22, 2021 |

Listas

You May Also Like

Estatísticas

Obras
3
Membros
1,528
Popularidade
#16,836
Avaliação
3.8
Críticas
25
ISBN
25
Línguas
4
Marcado como favorito
1

Tabelas & Gráficos