# Hypothesis and Scientific Method

Imagine if there is a layman observing a pond and a sea in the mild winter. He sees that the pond freezes while the sea remains liquid. He comes up with a hypothesis that 'Pond is smaller than the sea, that is why it froze. If the winter was harsher, both sea and pond would freeze.'

Although we now know that the salt concentration is the cause, but still, is his reasoning valid?

• What this example illustrates is the hazards of trying to generalize hastily from too little information.
– TKoL
Commented Jul 30 at 15:08
• We actually know that BOTH salt concentration AND size matter for a body of water freezing! Puddles freeze before ponds, and ponds before lakes and that is because of the high thermal conductivity of water, and the small thermal mass of the puddle. Commented Jul 30 at 17:48
• ... and salt water will indeed freeze if the temperature is low enough. Just how low it needs to be depends on the identities and concentrations of all solutes, but it might be only a few degrees colder than fresh water. Commented Jul 30 at 20:58
• "is his reasoning valid" There's no reasoning to validate here, IMO. A hypothesis is just that, if it doesn't contradict the known facts it's already "not invalid". Commented Jul 31 at 13:01
• It's a perfectly valid hypothesis which happens to be incorrect. Commented Jul 31 at 17:29

What you were (presumably) trying to present as an absurd conclusion is actually true, or at least it's part of the truth.

Some parts of the ocean do freeze, and area is indeed a significant factor in whether a body of water freezes. But there are also other factors, and for the entire ocean (or most of it, or just the surface) to freeze, that would require one hell of a winter, not the smallest issue of which would be that the ocean is spread across multiple hemispheres which have winters at different times of the year.

In any case, there are two problems with what you've presented:

### What you've presented is one specific expectation or comparison, not a model

In science, we're looking to create models (meaning some general rules for how things work, to use the term broadly). From those, we get expectations of what would happen in one particular case. Some expectation or comparison itself isn't your hypothesis.

A model may be that you need a lower temperature or you need that for a longer duration to freeze a greater volume of water. A more detailed model of thermodynamics may explain why this is the case.

One expectation from this may be that the ocean would freeze if exposed to lower temperatures than the pond (which is largely true), but there are other factors affecting the freezing point of the ocean (more on that below).

### What you've presented is only the first half of the scientific method

You observe, then you come up with a hypothesis to explain those observations. It is indeed valid to look at the unfrozen ocean and look at a frozen pond and come up with the hypothesis that bigger bodies of water take longer to freeze. This is where you've stopped, but there are more steps before it's reasonable to conclude that said hypothesis is actually true.

You need to also do experiments or gather additional data to test or verify your hypothesis.

You specifically need to try to eliminate other factors, which could include things like the salt content or chemical composition of the water, or whether the water is still or moving.

If you take different volumes of water with the same composition, in a still state, you will indeed find that one takes longer to freeze, so that would verify that hypothesis.

If you take the same volume of ocean water and pond water, or the same volume of moving water and still water, we'd see that those don't freeze at the same rate, so we'd come up with additional hypothesis to explain those, and indeed we did.

And all of those combine together, among other things, to explain why the entire ocean doesn't freeze, but the entirety of a small pond might.

Your example with the pond and the sea illustrates: It is not enough and rather useless to make just an ad-hoc hypothesis about the cause of an observed phenomenon.

One needs a hypothesis which explains the underlying mechanism. And this hypothesis should be testable.

In your example, pointing to the size of the pond and the sea is no substitute for an explanation. It does not aim at answering the question "why"?

Added as an answer to several comments: IMO it is necessary for a scientific hypothesis to propose an explanation how the mechanism behind the observed phenomena operates. Just adding further observations is not enough.

• They may not be an explanations, since they are merely hypotheses. But I most assuredly have the impression that the layman observing the pond and the sea is indeed aiming to answer the question "Why?". Commented Jul 31 at 1:50
• @DanielAsimov The layman from the example just adds a second observation: The pond is small, the sea is large. But this observation is no explanation. Commented Jul 31 at 4:30
• The relevant verb in my comment is "aiming". Commented Jul 31 at 14:44
• @DanielAsimov OK. Possibly he is aiming. But in any case, he is missing the target, an explanation :-) Commented Jul 31 at 14:58
• There's nothing wrong with the hypothesis that larger bodies of water freeze over slower than small ones. A hypothesis doesn't require a "why". Newton never proposed a 'why' for gravity, for example. Commented Jul 31 at 20:18

According to these definitions of hypothesis:

a supposition or proposed explanation made on the basis of limited evidence as a starting point for further investigation.

Philosophy :a proposition made as a basis for reasoning, without any assumption of its truth.

A valid hypothesis is not required to be correct. The scientific method is used to determine if the hypothesis is correct.

In your example, additional experiments to show the correlation between water body size and freezing points are required to determine the truth of the hypothesis.

• I think you are on the right track with this but you need to replace 'valid' with 'correct' here. A 'valid' hypothesis need not be 'correct'. An 'invalid' hypothesis is one that is non-falsifiable. Commented Jul 31 at 17:28
• @JimmyJames Thanks for improving my answer! I've updated to include your comments. Commented Jul 31 at 17:55

Although we now know that the salt concentration is the cause, but still, is his reasoning valid?

What does it mean to have "valid reasoning"? This is a question that goes back a long way through the history of philosophy and logic, and it has no simple answer. In fact, reason and critical thinking are difficult topics to understand and practice and take a lifetime to master. The poster child for reasoning in the popular media is Socrates with his gadfly-like incessant questioning, and among more recent philosophers, Kant and his critiques of pure and practical reasoning are infamously difficult to read and understand.

In the specific question you pose, the "scientific method", a strategy for reasoning about the physical world, is relevant because the unknowns in the physical world tend to be resolved by a unique epistemological process; it has been developed since the Pre-Socratic Greeks began shifting from gods to natural causes for explanations of the world around them. This sort of thinking goes by physicalism (SEP) and naturalism (SEP).

According to the scientific method (and to be accurate, there is more than one of them depending on the field of study), the reasoning process generally must include empirical evidence. From the article:

Empirical evidence for a proposition is evidence, i.e. what supports or counters this proposition, that is constituted by or accessible to sense experience or experimental procedure. Empirical evidence is of central importance to the sciences and plays a role in various other fields, like epistemology and law.

In the short example you provide, the thinker does not gather nor do they use empirical evidence, and therefore, their reasoning is not empirical, but only rational, and falls short of the standards science requires for good inference to best explanation. Science is in the business of providing reliable and accurate explanations of physical phenomena. Rational thinking is a necessary condition for good scientific reasoning, but it is not sufficient. Therefore, if asking whether the reasoning is valid, it would be fair to assess the reasoning as rational, but not empirical, and therefore not scientific because science requires gathering empirical data through testing. Valid scientific reasoning relies on that strategy.

As he has evidence to support this theory, and no evidence to the contrary, his reasoning is indeed perfectly valid.

He should test his hypothesis further for scientific purposes, but this does not detract from the validity of his reasoning.

is his reasoning valid

Sure. In the Scientific Method, any and all reasoning is valid as long as it is open to falsification.

So in this particular example, assuming there are no more known facts about seas and lakes to said observer, this is a perfectly fine thesis. It is open to further study - you can make a statistical collection and look at all the lakes and seas available to you. You can dig deep and try to find mechanistical explanations of why said correlation between size and freeziness is true or exlainable and so on and forth. It is falsifiable because you could do experiments where you make large and small bodies of water, controlling for any and all other conceivable aspects, and verify whether the small or lage body freezes first.

Even when, in this case, eventually the observer finds out about salt contents, the original hypothesis is still worth something. It may not be the ultimate reason why the first observation turned out as it did, but the volume of a body of water does matter in regards to whether it freezes over. As do many other aspects (like the air temperature, the impact of the sun, speeds of wind or currents, the salt content and so on and forth).

What would not be reasonable is if the observer came up with a reasoning that is not open for falsification. For example, they could jump to the conclusion that obviously there are small invisible water fairies which can only live in large seas (in a different dimension, as well) and not in small ponds (because reasons), and those fairies are what keeps the sea free of ice. This would be completely unfalsifiable, by construction, so it is not reasonable in the context of the Scientific Method.

Technically validity is a technical term within deductive reasoning that requires that if the premises were true the conclusion MUST also be true. Now science usually doesn't operate in that domain directly, because there's very little if any thing in nature that we know with such a certainty that we can conclude that something MUST be true/false. Science usually makes a assumptions and those come with error bars (even if they got pretty small over time).

So technically there could be a myriad of known or unknown reasons why despite of a harsher winter the ocean might still not freeze. Or idk there could be hot springs that are small, but won't freeze even in harsh winters. So in that technical sense those kind of reasoning is almost impossible to be valid.

Science is rather operating in the realm of inductive reasoning where you try to abstract a general rule/pattern from a series of observations. After which you apply a pseudo-deduction where you assume your rule to be certain and try to figure out what consequences that would entail and afterwards check if you actually see those, in order to further add credibility to your assumption.

So in such a colloquial sense of "valid" it might still be: ok-ish. You make an observation and abstract a pattern from it.

Though even with that milder sense of the word valid, there are things one could criticize about it. First of all 2 examples (pond and sea) is not a lot of data, so the generalization is pretty hasty. Also "smaller" and "harsher" are very qualitative terms, so in order to make that scientifically useful, you'd like to quantify that. What's the relation between size and freezing? Also there's a lack of definitions, like when is a lake considered frozen? When its surface is solid? When it's frozen all the way to the ground? The former would already imply that there are surface effects and the surface difference between ponds and the sea might be much smaller than the difference in water contained.

Also "harsher winter" is so impossibly broad that it might even support wrong assumptions. Like if you make it infinitely more cold pretty much all bodies of water will eventually freeze, but that doesn't mean that the amount of water or the surface area do play an important role in that.

So while it's ok-ish to conclude that size matters with regards to the freezing of water in smaller or larger accumulations, which you can support by observation. There is likely an argument to be made that about something something heat capacity, bigger surface to volume ratio, heat transport over that interface (surface) and so on. There's also a lot one could do to improve the hypothesis, to make it more quantitative and testable. So yeah it's not unreasonable to make that assumption given the data and the lack of data to the contrary, but it would be the starting point of further investigation and testing one's hypothesis and not be the end all be all of what there is to know about freezing of water.

• Oh, c'mon. You know what he meant by "harsher winter". Aren't there more substantive things to discuss? Commented Jul 31 at 1:54
• That it's colder for a longer period of time? But how cold and how long? Like these are qualitative comparisons that are hard to quantitatively test as you could always discard a counter example with "well it wasn't harsh enough!". Also a lot of the point that I discussed seem to have been picked up and improved in other answer (or the people came up with them independently). So what in particular would you like to have been discussed? Commented Jul 31 at 8:45

This is valid science

the simplest hypothesis that is falsifyable but has not been disproven is taken as the current valid theory

However, science doesn't consist of a single person! But from a plethoria of people, with constant interaction and discussions, who always try to disprove each other (in stark contrast to many other things, i.e. religion!). In fact, the best disprovers are Nobel Laureats, the scientific superstars.

It's easy for someone to come up with a counterexample. And the hypothesis is disproven.

_The mandatory sidenote is that the "most simple" solution should be accepted. There are always infinite possibilities that could also be, in addition to the explanation.

### The hard part

A new hypothesis always has to be able to explain not only a new phenomena, but also all the previously observed ones. That's why it gets harder with time to come up with new hypothesis. Conspiracy theories often fall into this fallacy: you don't need to explain one observation, you need to explain all of them.

As soon as some data about lakes, water and freezing is available, such a hypothesis will be easily rejected by the scientific community.

## The essence of hypothesis testing

This goes to the core of testing a hypothesis: we have a certain assumption about the world and test against a new assumption. Then we try to see, how likely it is that the result we obtain is actually compatible with our current world view. If not, we reject it. (omitting the stringent details).

## For all the claims of reasoning

Human reasoning is limited, and first and foremost, nature tells us what is. If you're on a "higher level" science, then yes, given that the lower level is sufficiently founded, you can find a reasoning. But on the lowest level (i.e. physics), we cannot just reason. Some things just are the way they are. And all we can do is capture the behavior mathematically.

(there are many famous discoveries where the reasoning failed, look for example the discovery of quantum mechanics, or the whole field of particle physics etc)

yes, many phenomena can be described as a consequence of other phenomena, but there are still fundamental laws of nature that we don't understand any better than the OPs example

• The most simple explanation should certainly be considered. But two things must be kept in mind: 1) The simplest explanation isn't always the right one, and 2) One person's simplest isn't necessarily another persons's simplest. Commented Jul 31 at 1:52
• The "most simple" solution should NOT be accepted, unless it is the "best fit" with the observed universe. For example, Newtonian Physics is simpler than Quantum Physics, but the latter is currently believed to be "more correct" Commented Jul 31 at 9:34
• @MikeB no, the other is not "believed to be more correct": you can easily design an experiment, where Newtonian Physics fails. Newtonian Physics is therefore wrong. Fullstop. (it's still a useful approximation!) Commented Jul 31 at 18:03
• @DanielAsimov 1) can you provide a counter-example? If it isn't the right one, it's not because we prefer more complicated versions but because it doesn't hold up in experiments and is therefore wrong 2) ish.There is a general consensus on what is simple, and while not at the detailed level sometimes, that's fine. Because it's a SE answer, not a book about the scientific method ;) The principle of "simplest" still applies Commented Jul 31 at 18:37
• "There is a general consensus on what is simple." I totally agree that this is the case sometimes. But not other times. Commented Aug 1 at 2:58