Thought experiment 3: The Gettier Problem

How the American philosopher Edmund Gettier’s argument complicates our understanding of what constitutes knowledge.

Despite the claims made by Donald Trump, more people came to witness Barack Obama’s inauguration in 2009 than bothered to turn up to his in 2017. Now consider the position of Jason, Hannah and Rachel in relation to this fact. Hannah has never thought about inaugurations and has no opinion about how many witnesses they attract; Jason believes Trump’s claim that his inauguration was watched by the largest audience ever; Rachel hasn’t seen any images from the inaugurations or read any reports about them. However, she has an entirely ungrounded theory that all US presidents whose surnames begins with T – Tyler, Taylor, Taft and Truman – are unpopular. On that basis alone, she is convinced that the Obama inauguration must have been better attended.

Do Hannah, Jason or Rachel “know” the truth? Well, obviously Hannah doesn’t. She doesn’t even have a belief about the facts. Jason can’t know it either, as he believes something false. Nor would we say Rachel “knows” the truth. Although she has a true belief, it is based on a spurious hypothesis.

The history is disputed, but at least by 1963, the Anglo-American philosophical community had come to a relatively settled understanding of what counted as “knowledge”. The tripartite analysis asserted that it was “justified true belief”. Hannah, Jason and Rachel each fail to meet at least one branch of this definition.

Then, that year, a little-known American philosopher, Edmund Gettier, wrote an article in the journal Analysis, citing imaginary cases that appeared to undermine this definition. Gettier taught at the University of Massachusetts Amherst and some years before he died in 2021, we had an email exchange. I was interested in him because his hugely influential article was only three pages long, and he published nothing else in his career. Had he been in a UK philosophy department he would have been a hopeless colleague as far as contributing to the Research Excellence Framework was concerned.

One case he discusses in the paper involves Smith and Jones: Smith and Jones both go for a job as a columnist. Smith has been told by the editor that Jones will get it. Smith sees Jones with a copy of the New Statesman in the pocket of his jacket. Smith believes that the person who’ll be offered the job is carrying a copy of the New Statesman. Does he know this? The problem arises because the editor changes his mind and Smith lands the job; Smith forgot that inside his own jacket he carries a copy of the New Statesman.

Other philosophers proposed their own variations. Suppose Smith is in a field and can see a sheep in the distance. Does Smith know there’s a sheep in the field? In fact, Smith is mistaking a large dog for a sheep, but there is a sheep grazing behind the dog. In both cases, Smith has a belief which is true and seems to be justified. Gettier triggered a vast literature. But non-Western philosophers had got there first. The eighth-century Buddhist writer Dharmottara imagined meat being cooked in a fire. Before it generates smoke, an observer sees what they think is smoke, but is a swarm of insects. The observer deduces that there’s a fire.

The typical reaction to these cases is that they are not instances of knowledge. There is a widely held intuition that there is a difference between “real” knowledge and “coincidental” knowledge. Indeed, lab experiments have shown that both primates and young children draw a distinction between agents who are knowledgeable and those who are ignorant – among the ignorant are those who get things right by chance. From around the age of six, children start to do something not seen in any other species: they recognise sophisticated states of belief (including, coincidentally, true belief).

One way a philosopher might respond to Gettier situations is this: knowledge is not merely justified true belief. It is justified true belief plus X. But what is this factor X? Numerous philosophers have hunted for it, but no non-contentious answer has yet been found. “There’s been a tremendous arms race,” the Canadian philosopher Jennifer Nagel tells me, “between analysis of knowledge and counter-examples to this analysis.”

Timothy Williamson, a revered Oxford logician, has proposed another approach. Knowledge, he argues, is not a function of several component parts. Instead, it is the primary concept, and justification and belief are best understood in relation to it, rather than vice versa. In other words, a belief is justified only if it is knowledge.

What Gettier reveals, says Nagel, is that “there’s a richness to our conception of knowledge that goes far beyond what we might have expected”. The hope had been to find a recipe to explain what knowledge was. It turns out “no such recipe can be produced. That’s a hard lesson. It’s provoked a deep anxiety about the fundamental relationship between knowing what’s going on in reality, and just happening to be right.”