# Are revealed preferences normative or positive notion?

I always thought from my understanding of the terms normative and positive that revealed preferences are a positive concept. For example, saying Anakin prefers grass to sand (i.e. $$U(g)\succ U(s)$$) is a positive statement.

My thinking is as follows:

1. It’s a value free statement

2. It’s empirically objectively testable statement. We can just make an experiment where we give Anakin choice of sitting in sandy room vs grassy room and observe preference.

3. Independent observers would agree on the observed outcome Anakin picking the grassy room.

However, I encountered an opinion that since all preferences derive from underlaying values they are all strictly normative and considering them positive even in the strict sense of revealed preference is a fallacy.

Is the opinion correct? If the opinion is not correct how would you refute it? Just by stating my points 1-3?

Also I tried to find some sources that would discuss this but google scholar just gives me results that are unrelated to this philosophical issues. I looked in Julian Ress’ philosophy of economics, and Hausmanns anthology but I could not find satisfactory answer there either. Are there any good references for this issue?

• I am not sure I understand the counter. What do underlying values have to do with this at all? "since all preferences derive from underlaying values they are all strictly normative" By the same logic when I send someone a romantic text message the physics describing the radio waves are normative as my underlying motivation for sending the message was normative. Or am I missing the point? – Giskard Dec 16 '19 at 19:06
• @Giskard that’s my interpretation as well... also the person (had this discussion with colleague from sociology department) argued further that this is also because the preferences are not just given but also developed by society... I don’t know what to think of it I think it’s wrong and the colleague told me if he is wrong I should prove him wrong... my worry is that due to imposter syndrome I honestly don’t know anymore if I am right or no and my major is not philosophy of economics (although I had few courses). I would like to know more opinions on the matter or some sources – 1muflon1 Dec 16 '19 at 19:13
• Sounds like your colleague is looking for a debate for the sake of debate. Unless you are looking for the same, I would let it go. – Giskard Dec 16 '19 at 19:17
• @Giskard well that’s easier said than done... but I guess you are right still if you know of some sources or some articles on this issue I would appreciate them. If not for the sake of debate just at this point I myself just need to know if this view is correct or not – 1muflon1 Dec 16 '19 at 19:19
• I cannot point you to sources other than the definition of the word normative. I have never heard it used like this. – Giskard Dec 16 '19 at 19:20

However, I encountered an opinion that since all preferences derive from underlaying values they are all strictly normative and considering them positive even in the strict sense of revealed preference is a fallacy.

There's a logical inconsistency here which might help see why it's not a good way of looking at the issue. What is the key element that is purportedly revealed by dividing the universe into "things normative" and "things positive"? I would suggest that it is determinism - matters of a positive nature correlate by way of fully deterministic, structurally-modelable processes that are discoverable through repeated experiment. Think physics - laws guide observables, observables reveal the laws, the laws can be used to predict future observables. If a tree falls in a forest and no one is around to hear it, physics tells us that it still makes a sound.

Matters of a normative nature correlate only to the extent that a system of "norms" has been set up to enforce them. Think ethics - morals guide behaviour, behaviour reveals morals, but can morals - absent a normative framework - be used to predict future behaviour? No - because morals themselves can change, because agents are capable of deception, and because agents may not even have perfect knowledge of their own preferences. Crucially, morals alone are no guaranteed predictor of unexamined human behaviour.

So, I'd argue that your friend's opinion is logically inconsistent because it is a contradiction: in order for it to be true that all preferences derive from underlying values, it must be true that values influence behaviour in a deterministic way according to underlying laws that are, in principle, discoverable through experiment. That's a pretty weighty claim, in my opinion not at all borne out by the preponderance of evidence - but if it were true, it would mean preferences are in fact positive!

Whether preferences are normative or positive then depends on the assumptions about the agent. We usually want them to be positive so that we can math about them, and so we make a whole bunch of assumptions about homo economicus that - if true - support that assumption. But of course, we know that homo economicus is a fiction. So we might suspect that our theory on preferences is at best a useful benchmark to provide context for understanding how people actually behave.

• I'm just highlighting the often-unstated fact that it's "turtles all the way down". Sure - you can observe and verify a person's consumption choices. But the mechanism that translates values into consumption choices is not observable, and in fact it's not totally clear that it exists at all. I'm not quite saying that dropping rationality makes revealed preferences normative - that's also a strong assertion. But, I am saying that if you claim revealed preferences are positive, you must be assuming rationality. Does that make sense? – heh Dec 17 '19 at 22:41
• As an example of how revealed preference might seem more normative in nature, check out the notion of "priming". You can look at the priming stimulus as an example of the "norm". It's not slam-dunk experimentally, and there are questions about the repeatability of some of the more famous experiments - but in a way, that's almost the point. :) – heh Dec 17 '19 at 22:46
• Yes, the whole point of (for example) giving a precise mathematical definition of "rational preferences" is so that we can proceed to overlay a preference relation with a utility function and get down to the business of doing mathematical economics with it. So whether someone says so explicitly or not, if they are relying on revealed preference as data, they are assuming the existence of a causal mechanism tying it to the underlying values - that's what the rationality assumptions do. It's stated so here (second paragraph): investopedia.com/terms/r/revealed-preference.asp – heh Dec 17 '19 at 22:56
• My preference (haha) is to begin at the beginning - Kahneman and Tversky's work on prospect theory. "Thinking Fast and Slow" is the quintessential popular-literature entry point, but there are lots of other good books - one I recently read is called "Risk: The Science of Politics and Fear". – heh Dec 17 '19 at 22:59
• thanks for all the help I appreciate all the explanations and source recommendations – 1muflon1 Dec 17 '19 at 23:01