The Five Virtues of a Hypothesis

[Gant]

The idea is what you come up with,,a hypothesis is the result you think willl happen from the idea:)

[Ophiolite]

You can have a lot of consensus on a hypothesis but it doesn't add to its validity. .......It's hard to accept a hypothesis on faith alone.

Just aman

 

I'll likely get accused of playing with words here, but surely we never accept a hypothesis on faith alone. Indeed, we never accept a hypothesis, full stop, period. We either reject it (having proven it invalid or found superior alternatives) or convert it (promote it) to a theory. The theory, if it is repeatedly substantiated across a spectrum of applicable circumstances, may eventually become a law.

Some persons, regrettably even some with scientific training, still confuse laws (and theories and hypotheses) with facts. The laws (and theories and hypotheses) are there to explain the facts.

[Guest mrcurious]

Quine (who died justa few years ago) was a distinguished American philosopher, perhaps one of a handful of "great American philosophers".

 

This general subject is "philosophy of science", and it's especially important for understanding science. The distinction between "theory" and "hypothesis" is very important; it's especially telling that the wackos who believe in creationism etc don't understand the differences, or refuse to acknowledge them. (I've tried to explain the differences, and the fascinating thing is, the anti-evolution crowd is very hostile to even listening to any of these ideas.)

 

BTW, iirc, Quine played an important role in WW II, in the effort to decrypt the material that the Germans had encrypted using the Enigma machine.

[lucaspa]

An hypothesis can vary from person to person. But surely an good hypothesis is based on at least some fact and asumption based on the facts.

When it is first proposed, a hypothesis comes from our imagination and doesn't have to be based on any fact at all. It is testing AFTER you form a hypothesis that establishes a connection to facts.

 

Now, hypotheses come in 3 forms, depending on the state of testing:

Untested.

Supported

Falsified.

 

But hypothesis can be wildly different if we take on intent as an ingredient in the building of the hypothesis even though facts dictate both hypothesis.

Example= The capitalist world is the ideal circumstance in which to invest capital for profit, so let the capitalist world thrive and let us make more profit.

Anti= The caspitalist ideal is to make more profit, at the expense of the poor, justice and the enviroment, so end capitalism.

These aren't scientific hypotheses, because both involve ethical judgments: "ideal" in the first and "justice" in the second.

 

Now, you can have competing hypotheses that both explain the data. For instance:

1. The fossils at Dinosaur National Park were deposited by a local flood.

2. The fossils at Dinosaur National Park were deposited by a world-wide flood.

 

Both would account for the data, since both involve a flood. The difference is the extent of that flood. To decide between the hypotheses, you need more data.

[sabbath]

A hypothesis usually states that something' date=' or in contrast nothing, will happen when you alter a single factor in an experiment. That nothing will happen (nothing out of the ordinary that is) is called the null hypothesis. That something (something out of the ordinary) happens is called the alternate hypothesis.

 

This means that an experiment tests two mutually exclusive hypotheses in order to determine whether or not a single factor produces an effect.[/quote']

 

This is how we state it in our research papers at school:

 

Ho(Null Hypothesis): There is no significant difference between the observed....

 

Ha(Alternate Hypothesis): There is a significant difference between the blah blah blah...

 

I wonder though if this is the only way? Is this the standard? We've been taught this, and we use this like a tradition.

 

To test a hypothesis do you always have to experiment on it in comparison to some standard?

 

What is the difference between a thesis and a hypothesis?

 

Thanks in advance!

[Glider]

Hypothesis are like food while cooking' date=' smell good, look good, but it cant be not eateable.

in other words... you can only do 3 things with a hypothesis.

1.- Prove that its true

2.- Prove that its wrong[/quote']

You can only disprove a hypothesis. You can never prove it, only generate results that support it.

[lucaspa]

An hypothesis can vary from person to person. But surely an good hypothesis is based on at least some fact and asumption based on the facts.

 

Not necessarily. You are talking about supported hypotheses. When a hypothesis is first made, it does not have to be based on anything at all; it is an imaginative construct.

 

However, the first step a person making a hypothesis should do is attempt to falsify it before he tells anyone else. This is where your "based on at least some fact" comes in. The hypothesis has already survived YOUR best efforts to falsify it before you tell anyone else. You have already tested it against facts to see if the hypothesis has any false consequences.

 

Now, the hypothesis can fail the testing. It is still a hypothesis, but it is now a falsified hypothesis.

 

So, you can see that hypotheses can be:

1. untested.

2. falsified

3. supported.

 

In these discussions, we tend to forget the first two and improperly assume that all hypotheses are supported hypotheses.

[lucaspa]

This general subject is "philosophy of science", and it's especially important for understanding science. The distinction between "theory" and "hypothesis" is very important; it's especially telling that the wackos who believe in creationism etc don't understand the differences, or refuse to acknowledge them. (I've tried to explain the differences, and the fascinating thing is, the anti-evolution crowd is very hostile to even listening to any of these ideas.)

 

There's not that much difference between hypotheses and theories. Both are statements about the physical universe. Theories tend to be more general statements and hypotheses more specific ones.

 

However, there is the misconception that certainty increases as you go from hypothesis to theory to law. IOW, that a theory is better supported than a hypothesis. This isn't true. There are very well supported hypotheses and some very speculative theories. And, of course, data can show any hypothesis OR theory to be wrong.

 

The creationist mistake is to use the colloquial usage of "theory" as "wild guess" in order to denigrate well-supported theories that they don't like.

[Genecks]

Use ABBYY fine reader next time Blike.

[lucaspa]

This is how we state it in our research papers at school:

 

Ho(Null Hypothesis): There is no significant difference between the observed....

 

Ha(Alternate Hypothesis): There is a significant difference between the blah blah blah...

 

I wonder though if this is the only way? Is this the standard? We've been taught this' date=' and we use this like a tradition.

 

To test a hypothesis do you always have to experiment on it in comparison to some standard?

 

What is the difference between a thesis and a hypothesis?

 

Thanks in advance![/quote']

 

This is confusing the statistical hypothesis with the research hypothesis. Two different things.

 

In statistics, you are trying to determine whether 2 sets of samples belong to the same overall population (null hypothesis or H0) or different populations (alternative hypothesis or H1). Population = all the possible examples.

 

For instance, you have a research hypothesis that a Zickel nail decreases healing time in a fracture of the neck of the femur (thigh bone).

 

You must test that by comparing people with fractures not treated with a Zickel nail and those people with fractures treated with a Zickel nail. The populations are 1) anyone who ever has had or ever will have a femoral neck fracture and 2) anyone who ever has been or ever will have a femoral neck fracture treated with a Zickel nail. You obviously can't get these populations, so you settle for samples of them. In an ideal world, you take the next 100 femoral neck fractures that come into the emergency room and you randomly treat 50 without the Zickel nail (controls) and 50 with the Zickel nail (experimentals). You then follow them with x-rays and other means and get a healing time for each patient.

 

If you plot the healing time vs number of people for each of these you will get a bell-shaped curve. The question is: do these bell-shaped curves represent different populations or do they represent one population and you just happened to get the appearance of 2 curves by chance?

 

THIS is your statistical hypothesis. The null hypothesis is that you got the 2 curves by chance and there is really one population. The alternate hypothesis is that there really was 2 populations.

 

So, you perform the statistics and we get two possibilities:

 

1. H0 (null hypothesis) results. This means that your research hypothesis is falsified. There really isn't a decrease in healing time with the Zickel nail.

 

2. H1 results. This means your research hypothesis is supported. You have failed to falsify the hypothesis -- despite your best efforts -- and now you can accept the hypothesis as (tentatively) true.

[richw9090]

Looking forward to the two remaining characteristics of hypotheses, Blike.

 

But I'll state what I think characterizes a valid hypothesis in science:

 

It is based on some set of observations (facts)

It is an attempt to explain or relate those facts in a systematic way.

It is framed in such a way as to be testable - that is, an observation can be imagined which, if obtained, would contradict the hypothesis.

It is productive - that is, it suggests further hypotheses (explanations) which can in turn be tested; each one that is confirmed lends further support to the originating hypothesis.

 

Rich

[Pymander]

You can have a lot of consensus on a hypothesis but it doesn't add to its validity. Common sense usually helps judge the propositions as blike put forward. To some minds though stuff makes sense to an individual first and he has to get consensus. Valid accepted arguments help support the case.

It's hard to accept a hypothesis on faith alone.

Just aman

Hypotheses are the basis of logical (deductive) arguments. Hypotheses state a property of reality as one perceives it. It represents an inductive step, arguing from the known to the unknown, a creative process computers are less suited to. Some examples of hypotheses are:

 

a + (b + c) = (a + b) + c states a property of addition.

 

"c, the speed of light, is constant relative to any observer" is a property of creation.

 

Newtonian mechanics was based on hypotheses which were only approximate by the factor 1/(1-v^2/c^2)^(1/2), which is insignificant when v << c.

 

We can not prove hypotheses, they abstract from complex reality by isolating variables. Without them, logical arguments or deductions must be infinitely derived or become circular (begging the question).

 

Hypotheses can only be formed from available evidence. Aristotle & Plato, Darwin, Hubble, and Freud did not have available the same evidence we have today. As evidence increases, hypotheses may need to be reexamined in new light, and arguments based on them must be abandoned. However, the status quo may, because of advantage, create massive inertia to such change. Frequently evidence is simply ignored and kept out of the light. This is supported by the history of revolutions in thought.

 

This very important basis for reasoning seem poorly understood, and the nature of 'proof'' is thus well exposed. It must always rests on unproveable hypotheses, which can be quite unconscious. "Common sense is that layer of prejudice laid down in the human mind before the age of 18." Albert Einstein.

 

Prejudice, axioms, and premises are other words for hypotheses. I hope this clears up a few matters, rather than confuses further.

Edited July 8, 2012 by Pymander

[MonDie]

This is a slight diversion, but it will be interesting as a creative idea about how to stimulate this sort of thinking in other people.

We already know that the process of neuronal death guides cognitive development.  If we can tie this to God's existence, we might get some productive discussions about truth and reality.  A pragmatistic understanding of knowledge that explains ideas as useful and functional---but ultimately subjective--tools in the interaction between the subject and his surroundings is not something that is easy to reconcile with the notion of consciousness without evolution via natural selection.  Selection among random variants, whether biological or cosmological, seems to be the only process by which functionality would emerge from randomness.  If you assume that randomness comes first and that cognition must involve functionally useful concepts, it is difficult to imagine cognition that is not the result of natural selection.

Will this taint Darwin's theory with filthy atheism?  Darwinian evolution =/= Atheism.  Evolution does explain life without providing an explanation for God, and it could potentially explain cognition through neuronal death without explaining God too.  And kinetic molecular theory doesn't explain God either.  However, if cognition has an explanation and God still does not, Huston we have a problem.  If we can discuss God's existence while discussing the cognition of information and facts, we could do society a service.  We will not taint Darwin's well established theory with atheism if we always distinguish it from speculative, philosophical theories about cognition.  Debating God's existence could be an antidote for misinformation if it becomes a debate about cognition in consciousness.

The sort of perspective I am describing is really not some tenuous theory but rather an uncommon way of describing something very common.  We always talk about thinking in the first-person and second-person.  This cognitive pragmatism is only the consequence of thinking about thinking in the third person and acknowledging the necessary relationship between perception and cognition (first person), which is the relationship between the perceivable world and our ideas about the world (third person).  The consequence of this third person perspective is that our ideas are not parts of the world but are tools with a functional, pragmatic relationship with the parts of the world they pertain to.  The more controversial question would be whether ideas that are functional tools are the result of an evolution-like process underpinned by neuronal death, but I think it is probably the case.  This should have consequences for where we should expect to find cognitive activity or the communication it produces, but also, because it produces analogies between the scientific method for a hypothesis and the trial testing of a tool, it has consequences for what sorts of popular ideas should or should not be verifiable by the populaces they infiltrate.

That is all for now.

Uh, happy - and hopefully safe - Thanksgiving. 👋😌

bye

[Markus Hanke]

I don’t think it is useful to resurrect a 17-year old thread, since none of the original participants is still active on this forum...

[John Cuthber]

17 hours ago, MonDie said:

If we can tie this to God's existence,

We can't.

 

[MonDie]

6 hours ago, John Cuthber said:

We can't.

In any case, I didn't want to discuss religion but to discuss how discussion of religion could be useful, so I will end with some comments on the opening poster's content.  Please note that I have virtually no philosophy studies.  It is a hobby.

---

Another advantage to this pragmatism or functionalism is the idea that a problem should be diagnosed before it is solved: If it isn't broke, don't fix it.  I will elaborate this later.  For now, the opening post repeatedly evoked my notion of precedent.  

 

[...] Acceptance of a hypothesis is of course like acceptance of any belief in that it demands rejection of whatever conflicts with it. The less rejection of prior beliefes required, the more plausable the hyothesis--other things being equal.

Often some hypothesis is available that conflicts with no prior beliefs. Thus we may attribute a click at the door to arrival of mail through the slot. [...]

The mail falling through the mail slot is not so parsimonious or consilient as it is having precedent--and moreover the phone dialing example seems to elaborate Blike's frequentist approach--.  The parsimony is in each instance of prior observation of the event having precedent, e.g. the mail slot: the parsimony is the corroboration from each sense that confirms what is happening, that mail is falling through the mail slot.  When you think it might be happening again, you are relying on precedent, not parsimony.  Blike's "modesty" elaborates the first complication of relying on precedent.

Virtue II: modesty:

One hypothesis is more modest than another if it is weaker in a logical sense; if it is implied by the other, without implying it. A hypothesis A is more modest than A and B as a joint hypothesis.[...]

This is the complication of categorizing similar events or things without violating parsimony.  For a human who is the product of evolution within an ecosystem of other evolved organisms and evolved structures, the Gestalt processes of perception are usually enough to efficiently identify recurrent things and iterative behaviors.  At least, that is what I personally arrived at.  Thus my diagnosis: Gestalt perception is necessary for the categorization of recurrent phenomena which is necessary for the notion of precedent, but it is evolved for the parts of a natural ecosystem and not scientific inquiry per se.  This is a potential diagnosis that could allow us to identify why people have problems reasoning about abstract scientific concepts.  Again, I am not an expert, but it is an example.

Virtue III: simplicity:

Where simplicity considerations become especially vivid is in drawing curves through plotted points on a graph. Consider the familiar practice of plotting measurements. [...]

[...] However many points we plot, there remain infinitely man curves that may be drawn through them. Whatever curve we draw represents our generalization from the data, our prediction of what bioling temperatures would be found at altitudes as the slimplest curve that passes through or reasonably close to all the plotted points. [...]

Now Blike was getting into the abstract, and that is why we are using a representation, a graph.  I have struggled with this a lot.  In my opinion, the idea of simplicity is more complicated because it so easily traverses the subjective/objective divide between the mental representation and the thing represented.   Sometimes the simple representation fails to capture the entire reality, because the mind must conform to the reality and not the other way around.  After the mind conforms, the corrected concept starts to seem more familiar and accessible to the subject, and the altered mental landscape means that subjective simplicity has aligned with the objective phenomena that is represented by it.  If this seems whacky, you might be right.  The same problem is precisely why I took issue with Blike's description of Conservatism.  Sometimes improper training of the mind, or improper reinforcement of the problem solving tendency, results in the increased accessibility of concepts despite the concepts' lack of parsimony or empirical utility.  Such an improperly trained mind will perceive Blike's "conservatism" where others will not.

I hope that was fun.  Peace.

I will revisit this thread to heck for interesting responses.