Surface tension is due to everybody trying to get into the water

Surely you are joking, Prof. Feynman? 1


Keith S. Taber


Photo of Richard Feynman, taken in 1984 © Tamiko Thiel (accessed from Wikipedia and shared under Creative Commons Attribution-Share Alike 3.0 Unported)


The late, great, Richard Feynman

Richard Feynman was special. Any one who wins the Nobel prize has to be pretty special, but physics laureate Feynman was even more remarkable as he was an exceptionally high achieving research physicist also known for his…teaching. No one gets a Nobel for being a good teacher, and it is often considered in Academia that teaching (that is, if one tries to give teaching the time and energy required to do it well – as students deserve) distracts from research to such an extent that it is rare to excel in both.

Feynman had something a lot of scientists do not not: great charisma. (That is no insult to fellow scientists – most plumbers and greengrocers and bus drivers and accountants and hairdressers do not – that is what makes it notable). He might be considered the Albert Einstein of the second half of the twentieth century, and because of that timescale we are lucky to have quality recordings of him talking and teaching in a way that could not have happened with previous generations. (A great shame in many cases: if perhaps a blessing with some – Isaac Newton's lectures were apparently avoided by most of his own students.)

Like many people, I find Feynman bewitching – he had a sparkle about him – almost a permanent mischievous twinkle in the eye – and an ability to somehow express the excitement of science (of working out why things are as they are) whilst being able to talk in ways that could be understood by people that lacked his expertise. That is perhaps one trait of a great teacher – being able to talk at the level of the audience, despite personally understanding at a higher, more complex and subtle, level.

That is by way of preamble – as I want to consider an explanation Feynman once offered of surface tension.


Screenshot of Richard Feynman explaining why water forms into drops.


Why does it rain in drops?

The extract I am discussing is taken from a 1983 BBC series of short episodes in a series called 'Fun to Imagine'. Although, at the time of writing, the episodes are "not currently available" from the BBC site, there is a compilation on YouTube. One of the topics Feynman discusses is the origin of surface tension – although he only introduces the technical term after explaining the phenomenon that water forms into droplets,

"you see a little drop of water, a tiny drop
And the atoms [sic, molecules] attract each other, they like to be next to each other
They want as many partners as they can get
Now the guys that are at the surface have only partners on one side
here, in the air on the other side, so they're trying to get in
And you can imagine…this teeming people, all moving very fast
all trying to have as many partners as possible and the guys at the edge are very unhappy and nervous and they keep pounding in
trying to get in, and that makes it a tight ball instead of a flat
and that's what, you know, surface tension
When you realise when you see how sometimes a water drop sits like this on a table then you start to imagine why it's like that
because everybody is trying to get into the water"

Richard Feynman speaking in 1983

Is this a good explanation?

Well, we might suggest Feynman makes a schoolchild error – water is not an atomic substance, but molecular. It does not contain discrete atoms, so he should be referring to the molecules attracting each other. But I do not think this is an error in the sense that Feynman was mistaken, simply that although the distinction is of great importance in chemistry, physicists sometimes use the term 'atom' generically to refer to the individual particles in a gas, for example. That might be unhelpful to a secondary school student studying for examinations, but if Feynman thought of his television audience for the recording as lay people, the general public, then perhaps the distinction between atoms (arguably a more familiar term in everyday discourse) and molecules would be considered an unhelpful detail? I am certainly prepared to give him that. I think it was the wrong choice, but not that Feynman was in error.

But what about the overall argument here. The 'atoms' want to have partners all around them 2 so they try to get inside the volume of the liquid. The overall effect of everyone, including these guys at the edge, trying to get inside the water is that it forms a sphere-like shape: "a tight ball instead of [something more] flat". Is that a convincing explanation – and is it a valid one?

What makes for a good explanation?

If anything is central to both science and science teaching, it is explanation.

"Explanation would seem to be central to the essence of science. A naïve view might claim that science discovers knowledge about the World, although it might be more accurate to suggest that science creates knowledge through the development of theories. The theories are used in turn to understand, predict and sometimes control the world, and in these activities, scientific explanations play the key role. We might consider theories and models to be the resources of science, but explanations to be the active processes through which theory is applied to contexts of interest…

An explanation is an answer to a 'why' question: but that in itself neither makes for a good explanation, nor for a scientific one. There is no simple answer to what does count as a good explanation, in science or elsewhere. Explanations have audiences, and to some extent, a good explanation is one that satisfied its audience – in other words it meets the explainee's purpose in seeking an explanation. Additionally, it has been known since at least Aristotle's time that we can talk of different kinds of causes, which suggests that many 'why questions' might have different types of acceptable responses, depending on the type of cause being sought."

Taber, 2007, p.159 [Download the chapter]

That passage is taken from a chapter where I described some activities used with secondary school students to help teach them about the nature of scientific explanation. (Read about the classroom activities here.) In that context, working with learners who were about 14 years of age, students were told that a good scientific explanation would be logical, and would draw upon scientific theory,

"pupils were told that scientific explanations needed to take into account logic and theory, i.e., that the explanation needs to be rational, and the explanation needs to draw upon accepted scientific ideas. As the notion of 'theory' is itself known to be difficult for students, they were also told that scientific theories are ideas about the world which are well supported by evidence; are internally consistent; and which usually fit with other accepted theories."

Taber, 2007, p.159 [Download the chapter]

Feynman's explanation is logical (if incomplete)

In that regard, Feynman's explanation can be considered logical, even if it omits (i.e., he takes as assumed) an important step* that is needed to explain the (approximately) spherical shape of the water drop.

If water quanticles (let's leave aside whether they are atoms or molecules) want to have many partners 2, and so try to get inside the volume, then we can understand* that the water drop will tend to the smallest surface area possible, so as few quanticles end up at the surface (with the tenuous air, rather than congregating water partners, on one side) where they will be nervous, and as many quanticles as possible are in the interior of the drop where they will be happy.

* The missing step is to state that a spherical drop will have a smaller surface area than any other shape with the same volume and so fewest quanticles at the surface. Perhaps Feynman assumed everyone would know/see that. Probably there is no such thing as a totally complete explanation.

So, is this a good explanation?

Explanations can have different purposes. Scientific explanations allow us to make effective predictions (and so often to control situations – the application of science through technology). But, in everyday life, explanations have a more subjective purpose ("explanations have audiences, and to some extent, a good explanation is one that satisfied its audience").

If, as a result of hearing Feynman's explanation, the viewers of the BBC televison programme

  • felt they now understood why sphere-like drops of water form, and
  • considered they had made sense of some science, and so
  • appreciated the value of science in explaining everyday phenomena,

then perhaps the explanation achieved its purpose?

Was Feynman's explanation scientific?

Of course, if I am being my usual pedantic self, I could point out that although Feynman's explanation was logical, that does not make it scientific unless it also drew upon accepted scientific principles. It was logical because the explicandum (what was to be explained – here, the drop shape) followed from the premise (i.e., if water quanticles want to have many partners, and act accordingly, then…)

But, in science, quanticles are not understood as sentient actors, but as inanimate entities that are not (and cannot be) aware of their situation and cannot act deliberately to work towards personal goals. Therefore, no matter how convincing someone may have found this explanation, it does not qualify as a scientific explanation as it is not based on accepted scientific principles (…or at least, not directly).

An anthropomorphic explanation

Feynman's explanation uses anthropomorphism, which from a scientific perspective makes it a pseudo-explanation. A pseudo-explanation takes the form of an explanation in that it is presented as if an answer to a why question, but does meet the requirements for a formal explanation (e.g., it does "not logically fit the phenomenon to be explained into a wider conceptual scheme", Taber & Watts, 2000.)

There are various kinds of pseudo-explanations such as tautology (circular explanations that rely on the conclusions as premises) and simply offering a label for the explicandum (e.g., water absorbs a lot of heat for a small change in its temperature because it has a high heat capacity – this is a kind of disguised tautology, as a 'high heat capacity' is a way of characterising something that absorbs a lot of heat for a small change in its temperature).

Read about pseudo-explanations

Anthropomorphism explains by assuming that the entities involved can be considered to be like people, and, so, to be sentient, have feelings and opinions and preferences, and be able to plan and carry out actions that are intended to being about desired consequences.

It relies on an analogy that may not be appropriate:

  • if people were in a situation like this, they are likely to behave in a certain way
  • if we treat these entities as if they were people then we might expect them to behave as people would, therefore…

It is an open question to what extent we can assume animals (chimpanzees, dogs, birds, etc.) can be considered to share aspects of human-like experiences, emotions, thoughts, etcetera. Perhaps it is reasonable to suggest a dog can be sad or a chimp can be jealous. It may not be stretching credibility to suggest that members of some species of animals want to be in large groups, like to be in large groups, and perhaps may even get nervous when isolated? However, it stretches credibility when we are told that viruses are clever or that a bacterium can be happy.

And, there is a pretty strong scientific consensus that at the level of individual molecules there is no possibility of emotions, opinions, desires, thoughts, or deliberate actions. Atoms do not want to fill their electron shells, and genes cannot be selfish, except in a figurative sense.

Read about anthropomorphism

So, in order to accept Feynman's explanation as valid, we would have to assume that the quanticles in water, water molecules,

  • like to be next to each other
  • want as many partners as they can get 2
  • can be unhappy and nervous
  • try to have as many partners as possible 2
  • try to get into the inside of the volume

So, to find this explanation convincing, we have to accept (contrary to science) that something like a water molecule is able to

  1. have desires and preferences,
  2. be aware of the extent to which is current situation matches its preferences, and,
  3. deliberately act to bring about desired outcomes

[Feynman does not explicitly state that the quanticles know about their situation (point 2), but clearly this is implied as otherwise they would have no reason to be nervous and unhappy, nor to act to bring about change.]

These requirements are clearly not met. A being with a central nervous system as complex as a human can meet these requirements, but there is no conceivable mechanism by which molecules can be considered sentient, or to be deliberate agents in the world.

So, even if Feynman's explanation of surface tension satisfies viewers of the recording (i.e., is is subjectively an effective explanation) it fails as an objective, scientific, explanation. Feynman may indeed have been a 'genius' (Gleick, 1994), and a great physicist, but his explanation here is invalid and simply fails as good science.

Now a reader may suspect I have gone after a 'straw man' target here. Surely, Feynman was speaking figuratively, and not literally. Of course he was, but figurative language cannot support a logical explanation, except through an analogy we suspect to hold.

Consider the following hypothetical claim and two possible consequences if the claim was true

ClaimConsequence 1Consequence 2
"I managed to get tickets for Toyah and Fripp's sold out concert in Bury St Edmunds, and these tickets are gold dust.""I could sell these tickets at quite a mark up""I could put a sample of these tickets in a mass spectrometer and would find they had an atomic mass of 197."

If the claim was literally true, then consequence 2 would follow. But of course, it is meant as a figurative claim, where 'gold dust' is a metaphor for something of high value because it is rare. So, actually consequence 1 might follow, but not consequence 2.

In the same way, if water particles do not have likes, and do not try to do things, Fenyman's argument seems to fall apart…

A teaching model?

Now I would not presume to know better than Richard Feynman, and I am pretty sure (i.e., about as certain as I can be of anything) that Feynman would not have fallen into the mistake of thinking that atoms or molecules actually act like humans and want things, or try to do things. He surely knew this was not a scientific explanation, but he clearly thought this was a useful way of explaining (to his audience) why water forms into a drop.

Now, I suggested above that Feynman's narrative account of the origin of surface tension "is not based on accepted scientific principles (…or at least, not directly)". But near the outset of this account Feynman states that the water particles "attract each other":

"the [molecules] attract each other, they like to be next to each other"

Feynman was not only a researcher, but a teacher, and teachers use teaching models. I think this is what Feynman is doing here:

"[according to science] the [molecules] attract each other [and we can think of this as if] they like to be next to each other"

Affinity in the sense of human experience is used as a kind of analogy for the affinity between water molecules (which leads to hydrogen bonding and dipole-dipole interactions). Once we model inter-molecular forces as being like attractions between people, we can extend the analogy in terms of how people feel when they do not get what they want, and how they respond by acting in ways that they hope will get them what they want.

Looked at this way, Feynman is doing something that good teachers often do when they judge a scientific model is too abstract, sophisticated, complex, subtle, for their audience; they simplify by substituting a teaching model which represents the scientific model in terms more familiar and accessible to the learners.

Read about making the unfamiliar familiar

From this perspective, Feynman's explanation may not have been a valid scientific explanation, but we might ask if it was an effective intermediate explanation set out in terms of a teaching model. That is, perhaps Feynman's explanation may have satisfied viewers, and also potentially acted as a possible foundation for building up to a more technical, scientifically acceptable explanation.

Teachers and other science communicators often use anthropomorphism as a way of offering accounts of complex scientific topics that will appeal and make sense to learners of a public audience.

Read about anthropomorphism in accounts of science

This can be effective to the extent that it engages learners, leaves the audience with a subjective sense of making sense of the science, and provides accounts that are often remembered later.

Of course that is not so helpful if the audience is studying a science course, and think they have learnt an account which will get them credit in formal examinations! I know from my own teaching career that learners often find anthropomorphic explanations readily come to mind, even when then they have been taught more technical accounts they are expected to apply when assessed.

In public science communication, then, anthropomorphic accounts may be valuable in offering people some sense of the science. But in formal education we need to be careful as even if anthropomorphism offers a useful way of getting learners familiar with some abstract topic (what might be called 'weak' anthropomorphism: Taber & Watts, 1996), we need to avoid them learning and committing to that metaphoric 'social' account thinking it is a valid scientific account ('strong' anthropomorphism).

Mapping Feynman's explanation

If we see Feynman as offering an analogy as a teaching model then we might seek to 'translate' his terms into more scientific concepts. He tells us that attraction is 'liking', and we can perhaps think of 'wanting' and being 'nervous' as indicating a higher (excited) energy state, 'pounding' as being subject to unbalanced forces, and 'trying to get in' as tending to evolving toward a lower energy configuration. At least, someone who already understood the scientific account could suggest such mappings. It seems unlikely any one who did not appreciate the science already could interpret it that way without a knowing and careful guide.

And like all anthropomorphic explanations, it 'suffers' from the very quality that it offers a narrative which is likely to be more easily understood, better related to, and more readily recalled, than the scientific account. This is why I have very mixed feelings about the use of anthropomorphism in formal science teaching, as even when it (a) does a great job of engaging learners and offering them some level of understanding, this may be at the cost of (b) offering an account which many students will find it hard to later let go of and progress beyond.

Screenshot of Richard Feynman explaining why water forms into drops.


As a good teacher, Feynman would know to pitch his teaching for particular audiences depending on their likely level of background knowledge. The explanation discussed here was not how Feynman taught about surface tension in his undergraduate classes at the California Institute of Technology (Feynman, Leighton & Sands, 1963). We can imagine that had he told students at Caltech that water formed into spherical drops because all the molecular guys are trying to get into the water, he might indeed had heard the retort: Surely you are joking, Prof. Feynman? 1


Work cited:

Notes:

1 My subtitle is a reference to the book 'Surely you're Joking Mr Feynman: Adventures of a Curious Character' in which Feynman tells anecdotes from his life.


2 Water was perhaps a poor example to choose as there is extensive hydrogen bonding in liquid water,

"I suspect even some experienced chemists may underestimate the extent of hydrogen bonding in water. According to one source…, in liquid water at the freezing point, the typical water molecule is at any time bonded by three or four hydrogen bonds – compared with the four bonds in the solid ice structure."

Taber, 2020, p.98

So, in Feynman's analogy, a water molecules does not become happy (lower energy state) when it is surrounded by as many other water molecules as possible, but when it is aligned with 3 or 4 other molecules to hydrogen bond, if only transiently. Without the hydrogen bonding, the drop would still be approximately spherical, but it would be smaller and denser as the molecules could get even closer together, but it would evaporate away more readily.