The argument that we cannot improve upon welfare has come up enough times that I feel like I need to argue for the natural alternative: the negative income tax (NIT).

There have been a number of proposals that I would consider to fit under this heading, most famously Milton Friedman’s proposal in “Capitalism and Freedom.” For the purposes of clarity, I define a NIT proposal to satisfy:

1. People who have no income receive something analogous to welfare, but smaller.

2. The amount of money that a poor person receives varies continuously with their income. That is, they gradually lose their welfare check as they make more and more money. Beyond a certain point, this decreasing payment becomes zero and then becomes the income tax.

If economic theory is remotely a good approximation of human behavior, it seems to be a no-brainer to me that this will lead to higher income among poor people than standard welfare. In the current system, welfare recipients have a disincentive to work- it is quite possible that with an especially shitty job, they will make less money in the workforce than they do on welfare. Thus, some people who would work if welfare did not exist will now not be working. By making payments gradually reduce as income grows, this disincentive becomes many times weaker. No matter what, having a job will increase your income. Simple, no?

Given that pure reasoning in an economic framework is not 100% reliable and might sometimes lead to thinking tinged with the libertarian bias inherent in economic theory (as much as I label myself as a libertarian, I feel that it is important to recognize that the intuition behind even very scientific models might have an ideological bias), I think it is good to see how this hits the data in the real world. The Earned Income Tax Credit functions in a manner similar to how a negative income tax would work. Here is an empirical paper on its effects http://harrisschool.uchicago.edu/About/publications/working-papers/pdf/wp_07_20.pdf .

Given how both theory and data seem to make the NIT seem like a better thought out choice than welfare, why is this not a political issue? How can people spend so much effort arguing back and forth about political ideologies when more good could come from simply improving the policies we have, regardless of the ideologies that brought them there in the first place?

1.2 A statement made in language does not directly express a truth about the world.

1.2.1 Since any formal definition of a word must use only other words, the process cannot “bottom out.” One must either accept some undefined terms or face infinite regress.

1.2.1.1 A formalized language, in which every term is either undefined or defined precisely through undefined terms, can thus make statements only about its undefined terms. A statement about a defined term reduces to a statement about undefined terms.

1.2.1.2 Undefined terms are in and of themselves meaningless. They are placeholders.

1.2.2 The assignment of truth values to statements in a formal language makes a claim as to how our undefined terms behave structurally.

1.2.2.1 By assigning truth values in a manner consistent with our some notion of deduction, we create a language which embodies a form of logic.

1.2.2.2 A language which embodies a form of logic which has not been contradicted by scientific investigation is quite useful. Such a language has the possibility of describing the world.

1.2.2.3 Any abstract structure is best described in a language which has a similar notion of deduction. Newtonian mechanics can be expressed in the language of calculus. Quantum mechanics can be expressed in the language of functional analysis. Quantum mechanics does not make Newtonian mechanics useless. Newtonian mechanics still makes coherent assertions within its language. The question of finding the natural language in which to speak of a certain topic is essential to any precise communication.

1.2.3 The properties a language guarantees about its undefined terms allows for meaning to be ascribed.

1.2.3.1 If one observes a system exhibiting properties that are mirrored by the undefined terms in a language, one can in some sense “perform the isomorphism” between the two and use the language to speak about the objects.

1.2.3.2 The meaning of a language is therefore not inherent to the language. It comes from our ability to use the language to mirror something else. The language does not mirror something else in and of itself.

1.2.3.3 Thus, we find a flaw with the picture theory of logic. One could say that logic is a picture, but it is wrong to say that it is a picture of the world. A picture is an object itself. By saying that one object or structure represents another, we must be saying that one can find an isomorphism between the two. It is only through perceiving the relations of a picture to the world that we can speak of the picture being any sort of representation. It is not inherent.

1.A fact in and of itself is not an atomic fact, just as a vector is not in and of itself a basis vector. To say that a proposition is a truth function of elementary propositions is thus a mistake. One must say instead that any collection of facts which provide axioms from which the world’s truths can be derived guarantees the truth values of a collection of statements, which we can call propositions.

1.1 The study of the possible mappings between different representations of the same logical space can thus be considered the fundamental research program of mathematical logic.

1.1.1 A deduction is a mapping with the property that if every statement in the mapping’s domain is true, then every statement in the mapping’s range must be true.

1.1.2 We should not view a logical space as merely a collection of statements. To define a logical space, we must also state exactly what qualifies as a deduction. Like Klein did with geometry, this approach naturally allows for the problem of classifying all possible logics.

1.1.2.1 In some sense, this is precisely Klein’s erlangen program. A geometric property, like a logical property, is defined merely by its invariance under a transformation in this approach. The words “geometric” and “logical” are undefined terms.

1.1.2.2 Frege’s program of deriving mathematics from logic, even if sucessful, would not have shown logic to be any more fundamental than any other part of mathematics. We could very well be able to derive logic from algebra as well as algebra from logic. One example of this in the real world is the application of Topos theory to studying logic.

1.1.3 A basis of atomic facts for a logical space combined with rules of deduction determines exactly what statements are in a logical space. Likewise, a basis of atomic facts combined with the entirety of the statements in a logical space gives us restrictions which our notion of deduction must satisfy.

1.1.4 Our experiences give us evidence that a collection of statements about the world is true. The image of these statements under all deductions constitutes precisely what we can know about the world.

1.1.4.1 Different definitions of deduction can lead to the extrapolation of different logical spaces, different worlds, from the same experiences. The determination of which definitions of deduction will lead to the extrapolation of worlds that are not contradicted by our future experiences is one of the tasks of scientific theory.