Scientists still enjoy a fairly high reputation in society as a whole (notwithstanding creationists and climate deniers). It is worth pausing to ask why scientists are still given credibility in this increasingly doubting age.
I’m pretty sure most people would say we deserve to be trusted because of the scientific method. You all know the 4 step version of the scientific method taught to every grade school child. But as I’ve pointed out while there is some combination of hypothesis, prediction and empirical test in good science, where those hypotheses come from is pretty unclear (its not true science proceeds by making up any guess you want). And things rarely follow a nice sequential order. There are usually a lot of backwards steps, circling around, and etc (nice post by Terry on this).
So if it isn’t the scientific method what is it? Predictions are certainly important. But there is plenty of science, especially in the early days of a field where a lot more observation than prediction is happening. And Baconian Empiricism (reality trumps pretty ideas) is certainly part of it, but plenty of theoreticians don’t go there.
Instead, I would argue that the one thing ALL scientists (and here I mean all types of scientists not just academics or people advancing the very edges of the field) have in common is we count. We measure. We put numbers on things. This is not exactly a new idea in science. For example:
BOSWELL. ‘Sir Alexander Dick tells me, that he remembers having a thousand people in a year to dine at his house: that is, reckoning each person as one, each time that he dined there.’
JOHNSON. ‘That, Sir, is about three a day.’
BOSWELL. ‘How your statement lessens the idea.’
JOHNSON. ‘That, Sir, is the good of counting. It brings every thing to a certainty, which before floated in the mind indefinitely.’
—-Life of Johnson; Apr. 18, 1783
Or
LORD KELVIN: “In physical science the first essential step in the direction of learning any subject is to find principles of numerical reckoning and practicable methods for measuring some quality connected with it. I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind; it may be the beginning of knowledge, but you have scarcely in your thoughts advanced to the state of Science, whatever the matter may be.”
— [PLA, vol. 1, “Electrical Units of Measurement”, 1883-05-03]
Or in ecology (and a bit more controversially):
ROBERT PETERS: “In the absence of a clear operational definition, different users of the term may develop independent, even inconsistent, definitions. In this way, the original conception grows by accretion of ‘conflation’ of meanings, until any single meaning of the concept appears restrictive and inappropriate. By that stage, the term represents a ‘non-concept’…”
— Robert Peters A Critique for Ecology 1991
It is probably worth noting that one can count and just be descriptive. And there are plenty of people critiquing descriptive science as stamp collecting. I don’t hold much truck with that though. Description (also known as natural history) is a necessary piece of science and some people only do description and are perfectly good scientists. Raunkiaer changed the field of botany and he didn’t do much beside count and describe. We certainly need some theoreticians and hypothesis makers too, but I don’t agree everybody has to do that. Moreover I am quite certain that a wildlife biologist measuring deer abundance or a botanist counting the number of endangered species on a plot of land are scientists. In academia we may call them technicians instead of professors, but I’m not trying to define what it means to be an academic. I’m trying to define what it means to be a scientist.
My bottom line – if you ask me what it means to be a scientists it means the following things going from most important at the top to least important at the bottom:
- Counting and measuring
- Predicting
- Social processes (peer skepticism)
- Scientific method in the traditional sense.
Some of you might argue that by the time you have counting and measuring plus predicting plus peer skepticism you have a good part of the scientific method, and I would say you’re right. But you’ve also lost a lot of mumbo jumbo that we don’t follow very often anyway.
I would also like to argue that by placing counting at the top of the priority list, one has hopefully sharpened our approach in the way Lord Kelvin suggests (“when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind; it may be the beginning of knowledge, but you have scarcely in your thoughts advanced to the state of Science”). In short if you CANNOT measure and count it, you should be a lot more worried about that than whether you have a hypothesis or not (Rob Peters did have a point even if he overstated it).
So what do you think. Can you be a scientist without measuring and counting? Can you measure and count without being a scientist?
Dynamic Ecology 
I think measuring necessary but not sufficient. So yes, you can count/measure and not be a scientist. But not the other way around.
In your prediction item, do you include “explanation” of existing data?
What else do you consider necessary to be sufficient to be a scientist?
I personally don’t require understanding of mechanism to make predictions but many do.
If I count the number of books on my bookshelves, am I a scientist? (Does it matter *what* we count?)
I think what we choose to count is the essential defining feature of a scientific field. Physicists study mass and energy. Ecologists study abundance, richness, nutrient fluxes, date of flowering. Notice that I always said count and measure. That is a much more sophisticated concept than simply enumerating the number of objects on my fingers.
So is counting the books on your shelf science? Maybe. But whether it is interesting science depends a lot on how and why you’re doing it.
I was reading Geographical Ecology on my commute in to work this morning (as one does) and thought of this post. MacAurthur writes:
“Numbers alone do not make science; it is relations between numbers that are needed.” He goes on to essentially say that just measuring or counting things doesn’t make science, but rather these counts or measures must be related to something else in order for it to become science.
In other words, observation is not enough. To be science, one must also use that observation in conjunction with other information to come to some conclusion. Science necessarily requires analysis, not just observation.
Hmmm. In the spirit of further exploring the boundaries of science…
It seems to me that an architect is doing lots of counting and measurement. She does a lot of analysis, both numeric and graphic, to test ideas and evaluate designs. Many different aspects of a design must fit together for it to work, and this takes a lot of integrative analysis and evaluation.
While most architects would not consider themselves scientists, there are may architecture faculty at university that consider the process of design a a form of research (in the academic sense).
It seems like a lot of natural history was not about counting, but more descriptive, even pictorial (i.e., Audubon’s illustrations).
I am more interested in the other end–counting but not science. What about Astrology? Lots of measurement there. Or Numerology in general.
Or what about Mathematics, particularly higher math that is all about symbolic logic. Is Computer Science a science? Or Cryptography–there can be a lot of counting in cryptography but that does not seem to be why it is a science.
On the other Hand I do agree, that to the extent what I do (landscape perception research and visual impact assessment) can be considered “science,” it is because of the measurement and counting.
As far math, it has been called the Queen of Science. But that is why I used “counting and measurement”. It is more about assessing and measuring the physical world.
If astrology measured not only the planetary movements (which I would call astronomy) but their predictions about people’s lives and their success rates, then I would call it a scientifically posed question (even if we all know it wouldn’t turn into a scientific field)
And what about the natural history question? Or a paleontologist trying to figure out how the bone she digs up fit together, or whether they represent a known species or something new? Not much counting, maybe a little measurement but little math. Science?
Personally, I think Audubon was doing natural history, not science. Which is NOT to say that all natural history is not science. A paleontologist is almost certainly doing more than digging something up and putting a skeleton together. They’re perhaps radio dating (or doing some other form of stratigraphy) and probably doing simple or complex morphometrics to compare to previous species. In short I don’t think the average Victorian naturalist was doing science.
I would say that both cryptography (at least developing it) and computer science are both well within the realm of science. I’ve heard computer science described as computer engineering, but I honestly can’t think of any criteria that by which we can effectively split engineering and science apart anyways. Astrology probably could be considered a science in its earliest incarnation. It only stopped being a science when its practitioners were unwilling to recognise any evidence disproving their ideas.
Also, effective description (as in natural history) is measurement in my eyes. If you can define something like “alternative leaves” or “bright red breast” in such a way that others can identify the same features unambiguously, it’s definitely a type of measurement.
I agree that measurement is nessecary and not sufficient though. I would say that scientists at least also have use their measurements to make generalizations about the world beyond what you measure. This would contrast with applying measurements for building rote designs (the day to day of carpentry for example) or applying measurements only to understand the thing measured (a tax collector may count wealth, but in early days they likely weren’t using this information to do more than say “you owe us x”).
I definitely think you’re on to something here. I don’t think there’s any way someone can be considered a scientist unless they are working with concepts that are at least measurable, and that they should have an idea of what a measurement of a given thing would look like. Reasoning about units was the first thing I taught as a TA in theoretical ecology. Unit analysis and the role measurements play in science doesn’t often get emphasized in ecology classes (except in labs) but I think that’s for the same reason that fish probably don’t talk much about the importance of water.
I would say you can measure w/out being a scientist; I worked building roofs for a summer, and while it required a lot of measurements, it definitely wasn’t science. Further, governments have been measuring tax income for hundreds of years w/out any scientific understanding.
When a field starts taking effective measurements may be a good point to mark when it is becoming science though; I would say you can definitely mark the transition of macroeconomics (for instance) from philosophy to a science around the time when the national accounts were developed, even though I don’t think they were developed primarily to advance science.
Interesting analogies with economics. It makes a lot of sense to me.
It seems we might be on the same page in thinking that serious science starts with a conversation about what is worth measuring, how it is measured, and what we can or cannot measure.
I know this is about counting but I really like the inclusion of #3 (peer skepticism) to your list of the hallmarks of science. Especially given subfields in which some concept or method is so hardened that there is little within-group peer skepticism. I don’t quite mean zombi-ideas, but just a concept, view-of-the-way-the-world-works, or analytic tool that isn’t what the peer group thinks it is. Gelman’s blog has focused lately on social psychologists because this field is such low hanging fruit but I suspect there is a bit of this in all subfields. I think this is the advantage of post-hoc review (no I’m no advocating the abandonment of traditional peer review) and why science as a whole would benefit from a richer set of post-hoc review mechanisms.
You asked – I have done a complete PhD in Mechanical Engineering and Biology (Biomimetics) without ANY statistics at all (i.e. no counting) and yes you can measure without measuring by sliding your speciem on a slide and putting it under a confocal microscope.
Yours
Bruce
What do you mean by measure here? The dictionary definition is “ascertain the size, amount, or degree of (something) by using an instrument or device marked in standard units or by comparing it with an object of known size.” So we have an instrument, but I assume you are evaluating some sort of pattern you see in the specimen. Maybe “measurement” it not right concept; maybe it is systematic description, which could include measurement?
One way to test the edges would be to think about whether and when researchers using qualitative data are doing science. and maybe why a good investigative journalist is not doing science (or is she?).
This is fun. Is this they type of discussion that a post from a couple of days ago was lementing? If so, why is this post working?
Correct me if I am wrong, but measurement is for transfering of data. If one finds a way of doing somethign without need to measure and yet it fits…..?
Systematic description is a good phrase. And I like your question about an investigative journalist. I think it is a great example in fact. An investigative journalist and a sociologist could study the same issue (e.g. causes of gun violence in US) but I would say the sociologist is doing science and the journalist isn’t. And the quantification has a lot to do with it.
I feel honesty is the most important characteristic of being a scientist. A scientist must be willing to reject a hypothesis when the data do not support it. I am not certain this is always the case.
I would put Observation & Awareness as number 1…and by ‘observation’ I mean real ‘eyes wide open’ noticing things going on around you. Anyone can count & measure, but knowing what will be interesting/useful to count & measure comes with awareness of the system.
You also need to know how to put a jigsaw puzzle together! Counts & measures don’t mean anything if you don’t understand, or can’t explain to anyone, how those data fit into the system.
As long as “observation and awareness” is interpreted sufficiently broadly, I guess. But if it means “observation and awareness of nature”, meaning that you have to be outside observing nature and taking those observations as your starting point when deciding what questions to ask and how to answer them, then sorry, but I’m not going to go along with that. Because it would define me and people like me (such as Ben Bolker) out of ecology, or out of science in general:
I wasn’t restricting it to nature and I’m not defining you out of ecology. I was referring to being aware of your scientific jigsaw and understanding what science you’re doing first, as opposed to having your head in the sand and just counting & measuring for the sake of it. Is there a point to measuring sound waves if you have absolutely no idea where the sound is emanating, what’s making the sound, and what system it’s audible in?
Ok, thx. People have tried to define me and others out of ecology in the past (presumably without realizing they were doing so), so I tend to be a bit sensitive about it. 🙂
Hi Brian:
I agree with you that the standard model of the scientific method is a flawed way of thinking about science, not least because we know from 50 years of experiments in cognitive psychology (going back to the early ones that Kuhn talked about) that observations are not model-independent, so the idea of starting with some sort of naked, unfiltered observation and then formulating an explanation is somewhat fantastical.
However, I think the discussion thus far has omitted the most important distinguishing feature of scientific knowledge: reproducibility (unless I missed it in the comments somewhere). Empirically, experiments and observations can be repeated. In statistical modelling, analyses can be repeated. In theoretical modelling, derivations can be checked. Results that can’t be reproduced are typically (eventually) discarded. Results that are reproduced become an accepted part of the scientific body of knowledge.
This captures all forms of science, I think, including non-counting kinds like descriptive taxonomy and so forth. It works for contextual sciences, such as ecology, and also for less-contextual sciences, like particle physics. I think it even works for things like the Big Bang, since theoretical predictions about how the universe should look now can be tested, as can the derivation of those predictions. There may of course be predictions we can make that are testable in principle but that we can’t test yet, given our capabilities (string theory comes to mind), and so I think it is fair to consider these scientific theories, but theories that lack much (or any) empirical support.
I’m not sure I would go so far as to say that this is a sufficient condition for something to be considered scientific, but I will tentatively venture that it is a necessary one.
Cheers,
Sean
what if the hypothesis is a random walk or contingency? These results cannot be reproduced (although certain parameters can, say the variance of a random walk) but they form important hypotheses in many areas of science.
Interesting – I totally missed reproduceability (and it hasn’t come up in the comments). But you are right it is an essential feature.
I’ll just leave this here. 🙂
Hi Jeremy:
By reproducibility, I definitely don’t mean just replication. (My entire PhD was on an unreplicated and unreplicable gradient in oceanographic conditions!) While some experiments (such as Brian’s Biosphere example) are often not replicated, in principle such experiments can be repeated. Likewise, observations on unreplicated natural gradients can be repeated.
I concede that causal explanations for such observations (or similar large-scale temporal patterns, like a long-term macroevolutionary trend for some group of organisms) must be probed by asking whether they operate in the predicted way along other gradients that have a mixture of shared and unshared features with the one that was first examined, and/or by identifying other predictions that should hold if the causal explanation is valid. All of these things involve further empirical investigation that can contextualize, constrain, or overturn the inferences drawn from the initial set of observations. So perhaps “reproducibility” carries too narrow a connotation, but hopefully you get the idea.
Contrast this with at least some religious knowledge claims (for instance, about the occurrence of miracles), which, because they involve an arbitrary suspension of the ordinary “laws” of nature, must be accepted or rejected based wholly on one’s willingness to rely on the testimony of others.
A corollary of reproducibility, as I’m using it here, is that scientific knowledge of the natural world is provisional — it can always be changed in light of future findings. Knowledge about the world obtained by special revelation (for instance) lacks this important feature of provisionality.
Cheers,
Sean
Interested to see this thread digging into what philosophers of science call the “demarcation problem”–how to distinguish science from pseudoscience. Popper’s idea of falsifiability, for instance, was originally proposed as a solution to the demarcation problem (at least that’s my understanding). Science is falsifiable, pseudoscience isn’t. Philosophers of science haven’t found any solution to the demarcation problem fully satisfactory (again, that’s my understanding). I’ve touched on this a little bit, in an old book review:
How is culinary ‘art’ not a science? There is a lot of calculation – counting and measurement to a very nuanced degree, social evaluation, certainly, and reproducibility (or replicability) ad infinitum with a high degree of predictability.
Personally, I am increasingly seeing the thing I took for granted such that I didn’t even mention it is an attempt to answer unknown questions. Within that context, I think my list of measurement, prediction, peer skepticism stands as a definition of science I would defend (although reprodceability is a good addition).
To answer unknown questions – in other words, to expand knowledge. We go back to where it began … Scientia. If this is fundamental or the framework within which to place your definition, the concept or notion of ‘science’ would embrace or accommodate many more disciplines than what it has come to mean today? And, if I may venture to ask, scientists are less an exclusive breed than imagined?
To answer questions + the framing of this post (quantitative, predictive and skeptical).
I don’t think scientists are particularly exclusive. That is one of my points.
Pingback: Observation as an essential ecology skill » EC0L0GY B1TS
Pingback: I skype therefore I am a scientist? | Dynamic Ecology