January 6, 2014, by Brigitte Nerlich


There has recently been some discussion in climate change circles about climate sensitivity and predictions of warming trends about which I will not talk, as I have no expertise in those fields. However, as I am nevertheless trying to keep up-to-date, I recently read a reblogged blog post about just such issues on And then there’s physics. A little debate about ‘certainty’ in the comments caught my attention.

Conversations about certainty in science

One commenter pointed out that we cannot know for certain what will happen in the future. This provoked the following reply by the blog owner: “Of course we can never know what will happen in the future. Science isn’t about certainty, it’s about understanding what might happen and assigning confidence to the various possibilities.” This is in turn elicited the following response from the commenter: “I disagree. Science IS about certainty. It’s about having a theory which matches observations and being able to produce repeatable experiments. Speculating on what might happen and assigning confidence is in the realm of tea leaves and tarot cards.” Another commenter then said: “And all these years I thought the health professionals I and my family consulted were medical scientists when they gave us the probabilities of successful operations […]. In future, we will be demanding CERTAINTY from these practitioners or they’ll be relegated to the level of psychics, mediums, witchdoctors and lawyers!” Yet another commenter chipped in with a comment on (natural) ‘laws’ and pointed out: “Sometimes, all you can hope for are law-like regularities. There are also explanations, and observations, and theories, and hypotheses, and models, et cetera, none of which are certain.” These are all topics about which one could write whole books!

After reading this, I moaned to some friends: “But what does certainty actually MEAN in science?” – and got this answer from a physics friend who laughingly said: “It’s when you’re getting over-confident with your results”. Which sort of sums things up quite nicely for me, but there’s certainly more to it than that. I am not a natural scientist though, so I’ll try to tackle the ‘more to it’ question from my own linguistic perspective.

Languages of certainty and uncertainty

In the past I have done a bit of digging into the ordinary language use of ‘uncertainty’, which can contribute to confusion, as the meanings of the word in scientific jargon and in ordinary language are quite different. I tried to dig a bit deeper this time and used an archive of ordinary language use which collects data from the web. I wanted to see whether rummaging around here might confirm my previous results and shed some light on the use of certainty as well.

I used WebCorpLSE which is a ‘Synchronic [contemporary] English Web Corpus’ consisting of 457,713,650 words from web-extracted texts. It covers the period 2000 to 2010 (there is also WebCorpLive, which I haven’t used). A search for ‘certainty’ yields 4711 hits; for ‘uncertainty’ 8071. This seems to show that we talk much more (worry much more) about uncertainty than certainty, and increasingly so it seems, after the second world war (see graph from Google Ngram Viewer that I used as a featured image, where the blue line is ‘certainty’ and the red is ‘uncertainty’).

As Ludwig Wittgenstein said, to understand the meaning of a word, you have to understand its use in a language. And as the linguist John Rupert Firth said: “You shall know a word by the company it keeps” (Firth, J. R. 1957: 11) So let’s see what company my two words keep. For this I asked WebCorp to look at words used next to or adjacent to the words certainty and uncertainty (in this case only preceding them).


Some of the most frequently used words adjacent to certainty (after function words such as ‘a’, ‘the’, ‘with’, which I only looked at very cursorily) are: ‘great(er)’, ‘absolute’ and ‘100%’ (and various other percentages around 90%; the IPCC‘s definition of 90% = very likely gets a mention), followed by ‘large’, ‘more’, ‘near’, …, and most of all, ‘no’ certainty! There is talk of a need ‘for’ certainty and a quest ‘for’ certainty, at least ‘of’ a certain degree, level or kind.

Certainty is used together with ‘absolute’, ‘dogmatic’, ‘infallible’, ‘transcendental’ and ‘dead’. The word is linked by ‘and’ to knowledge (‘and’ certainty), stability, power, clarity, perspicuity, measurability, confidence, ease, simplicity, predictability, speed and arrogance, for example.

There is ‘moral’, ‘legal’, ‘regulatory’, ‘medical’ and ‘clinical’, but also, of course, ‘scientific’, ‘mathematical’ and ‘statistical’ certainty. However, there are many instances of ‘lack of’ and ‘absence of’ (‘full’) (scientific) certainty. Overall, the word certainty seems to be used in some scientific contexts, but mainly with caution. Overstating certainty seems to be associated with dogma and arrogance.


Many adjacent words seem to indicate that we speculate more about the size of uncertainty than certainty. Uncertainty is preceded by  ‘considerable’, ‘substantial’, ‘continuing’, ‘great(er)’, ‘major’, ‘high’, ‘increased’, ‘growing’, ‘much’ and many more. There is also talk of ‘creating’ uncertainty, as well as, in a few instances, of ‘manufacturing’ uncertainty (in the context of climate change). Occurrences of ‘no’ are much rarer with ‘uncertainty’ than ‘certainty’.

Some adjacent words confirm my previous research into ordinary language use, where uncertainty seems to be linked to quite a bit of existential angst. In the company of ‘and’, we find talk about fear, anxiety, stress, confusion and so on. ‘And’ also links ‘risk’ to uncertainty in a number of instances. Risk and uncertainty seem to be quite a couple. There are, of course, also expressions of worry about ‘economic’, ‘financial’, ‘market’ and ‘future’ uncertainty.

Most interestingly perhaps, uncertainty has a much clearer scientific signature than certainty. We find words like ‘experimental’, ‘inherent’, ‘statistical’, ‘parametric’, ‘quantitative’, ‘measurement’, ‘model’, ‘modelling’, ‘significant’, and ‘reduce’, for example. Heisenberg’s uncertainty principle makes various appearances (as well as, amongst others, Gaussian, Hamiltonian, and Knightian uncertainty).

Alongside science we also find some uses of uncertainty in the context of politics, such as ‘political’ and ‘policy’ uncertainty, as well as decision making ‘under’ uncertainty.

What does this mean?

These very rough findings seem to indicate that when people talk about certainty the word takes on an almost religious meaning. This is rather different to how the word is used in scientific contexts, as science deals with probabilities or likelihoods, not absolutes. When the word certainty is used in more scientific contexts, words of caution accompany it.

Uncertainty is used in two contexts: in everyday talk and in science talk. In everyday talk, the word is often used together with fear, confusion, doubt or anxiety. When used in scientific contexts, the meaning of the word seems to agree with its established meanings in science. As Stuart Firestein pointed out in his 2012 book Ignorance: “Being a scientist requires having faith in uncertainty, finding pleasure in mystery, and learning to cultivate doubt. There is no surer way to screw up an experiment than to be certain of its outcome.” No fear or anxiety here! So there is quite a gulf between everyday talk and science talk which can lead to misunderstandings.

Overall, these snippets of information about how we talk about certainty and uncertainty on the web show how easy it can be to talk past each other, intentionally or unintentionally. The words certainty and uncertainty have multiple meanings and we have to be careful how we use them and for what we use them, especially in the context of climate change. In a sense, the IPCC has been trying to learn this lesson and has devised a whole new language to express levels of confidence and uncertainty about climate change in its reports. This does not mean, of course, that the word certainty is not used in debates about the IPCC – far from it. But that would be a topic for another post.

PS A more philosophically inspired essay on science, politics and certainty can be found here

Posted in Climate ChangeLanguageScience