Introduction

This paper is mainly concerned with the issue of epistemology in economics, i.e., the relation of ideas with reality¹. The central question is: does the near-zero marginal cost of production of new technology is really bringing the demise of capitalism? From the perspective of neoclassical economic theory, the short answer to this question is “no”. But answering too quickly risks missing the underlying epistemological issue at stake. Arguably, the epistemological issue is not strictly within the limit of the economic valuation of the factors of production. It is precisely when one dismisses the means of economic valuation that epistemological issues arise. The risk is that dubious views vaguely assimilated to economics, may take center stage in formulating economic policy² applied to new technologies. The significance of addressing the epistemology of the value of the factors of production points directly to the value of economic calculation in public decisions. To understand its significance, one needs to appreciate:

First, today’s professional economists do not consider the epistemological issues of economics in their core activities. In other words, economists are unwilling to engage in defense of economic calculation since this task falls outside their strict professional duties. This leaves a chance for the economic illiterates on price valuations to dominate the market of ideas.

Second, the key role of economics remains (in its most simplified version) to valuate the output of productive actors at all possible levels of analysis. If economics is to serve policy objectives, prices and quantities remain one of the most transparent means for economics to inform politics. Accordingly, we address the epistemological issue to know what it means to undermine the valuation tools of economics (here, the costs of production factors) to generate a public debate on new technologies.

To disentangle the issue between unidentified disciplinary limits, as Rifkin’s attack on the role of “near-zero marginal cost of production” in information technology, we use two methods of reasoning sequentially. The first one is the recognition that the marginal cost of production falls under the valuation of the factors of production. Accordingly, to assess the marginal cost of production, textbook economic theory specifies how fixed, variable, and marginal costs operate. The second is hermeneutics, i.e., scrutinizing all referencing, information sources, and other available documentation Rifkin used to set the stage for capitalism’s self-destruction. The following part presents Rifkin’s starting point on the marginal cost of new technology production.

1. Rifkin’s narrative of capitalism self-destruction

Ogden (2014) correctly reviews the main arguments in Rifkin’s book, which claim that the sharing economy based on technological communities will overthrow corporate capitalism. Rifkin points out that “collaborative common”, “sharing economy”, “internet of things,” and “new energy development” are available at a minimal cost reduce corporate profit margin. His theory presumes that new sharing technology will be more efficient than a corporation to deal with capital efficiency and coordinate the factors of production. Ultimately, he wants to show that technological communities’ low cost of production will presumably reduce firms’ profit margin so much that it will ultimately end up destroying capitalism itself.

Ogden perceived right away Rifkin’s reasoning fallacies³. He also recognizes he may have failed to identify what kind of game Rifkin is really playing by declaring the demise of capitalism. This paper aims to identify that fallacy. We will show that Rifkin is setting up an economic concept - the so-called “near-zero marginal cost” of the sharing economy - to do a non-economic comment on new collaborative technology. How does he proceed to do that?

Rifkin uses this expression ambiguously to refer to the cost of production and the distribution of information technology. On the consumption side, the near-zero marginal cost refers to the marginal cost of reproducing one unit of technology. If one unit is sold to 100 clients, its marginal cost of reproduction is 0,01. With the internet, that same unit may reach 100.000 users, bringing the marginal cost of reproduction to 0,0001. On the production side, however, Rifkin (2014, p. 9) discounts both fixed and variable costs, rendering the calculation of the marginal cost of production impossible:

He dismisses the fixed cost (here loan, rent of working space independent to output):

(…) the cost of actually producing each additional unit-if fixed costs are not counted-becomes essentially zero, making the product nearly free.

He discounts variable costs (here, the marketing, distribution, and labor dependent on output):

The cost of marketing and distributing each copy is nearly free. The only cost is the amount of time consumed by creating the product and the cost of computing and connecting online. An e-book can be produced and distributed at near-zero marginal cost.

From the production side of the new technology, economists know that the marginal cost implies both fixed and variable costs of production. So, the cost of producing a new unit of online journalism, a cellphone, a 3D printing object, online education, music on a streaming platform etc., cannot be close to zero. Rifkin is not strictly using Marxist economics either. In a planned economy, the production manager would impose the price of the product to match exactly its marginal cost of production, so that there would be no profit (SCHUMPETER, 1942, p. 176; HAYEK, 2015). In a free market economy, the entrepreneur chooses a price proportional to all costs of production (fixed and variable costs) that yield maximum profit within the constraints of market prices. One important remark is that Rifkin’s dismissal of fixed and variable costs renders it impossible to put a price on the factors of production. In this sense, Rifkin operates within the socialist commonwealth, whereby money does not play a role in the economic calculation (MISES, 1990, p. 21).

We suggest that, since Rifkin is not using the economic valuation of the means of production, namely fixed, variable, and marginal costs, he operates outside economics. The solution we found is that Rifkin refers to “the marginal cost of production” semantically, as an instance of a historical materialism’s interpretation (MARX, 1976, p. 1) of the vanishing of profit and the subsequent demise of capitalism. Let us test this suggestion on his own writings. The goal is to see if Rifkin operates within the Marxist rhetoric⁴ and not under the neoclassical framework. Since our interpretation seems plausible, why should we bother to verify it further? There is one good reason for doing that. If Rifkin does not operate within economics, the reader of his book will have no means to understand the dynamics animating new technology.

To appreciate the difference between the use of the same expression, one operating in textbook economic theory and the other in Marxist ideology, we need to show how each of them operates within different language games⁵.

2. A note on method: a Wittgensteinian examination of language

The distinction between the economic expression “marginal cost of production” of new technology, and Rifkin’s Marxist misuse of it, is shown by common-sense evidence⁶, not empirics. This is why it calls for a method appropriate to that kind of problem. The “ordinary language philosophy” (HACKER, 1996; PASSMORE, 1957; MCGINN, 1997; RORTY, 1967) seeks precisely to distinguish between the use and misuse of given expressions. Philosophers have used this method of language scrutiny to solve philosophical puzzles⁷.

Ordinary language philosophy is used to identify, clarify, and even ‘dissolve’ confusions. It is particularly effective to deal with problems created within our language, by means of our language. We can demonstrate this method by showing that the problem Rifkin is dealing with is not an empirical problem of cost calculation, but a problem of misuse of economic expressions. To answer the epistemological question concerning the link of some ideas to economic reality, we need to distinguish sharply between the cost valuation of new technology, and the diffusion of unverified assumptions about it.

To make this distinction, we use the grammatical investigation of language to distinguish which language game one expression belongs to. The method is based upon a scrutiny of language use. In this case, it allows differentiating between the principles for calculating production costs and Rifkin’s use of it. That difference is not a matter of hypotheses in natural science. In other words, no scientific method can be applied to settle the difference between “neoclassical theorizing” and Rifkin’s Marxist rhetoric. The only way to do that correctly is to engage in what Wittgenstein calls a clarification of the misunderstanding of the logic of our language (WITTGENSTEIN, 1958, p. 93):

[Those misunderstandings are] not empirical problems but are misunderstandings that are solved […] by looking into the workings of our language, and that in such a way as to make use and recognize those workings: in despite of an urge to misunderstand them. (WITTGENSTEIN, 1958, p. 109)

On the surface, Rifkin seems to be dealing with the cost function in economic theory. The goal of the ordinary language philosophy is hermeneutical, i.e., a close study of the use of words in a text. It is the only method that helps the reader clearly realize that Rifkin operates within the realm of a “narrative of social destruction”. To be more scholarly precise, he operates within the Marxist materialist narrative of the demise of capitalism. Here, Rifkin’s use of the expression “marginal cost” is nothing more than the absurd Marxist presupposition that one can deal with cost evaluation outside economic calculation. As Mannheim said (1936, p. 89).

the danger in presuppositions (…) lies in the fact that an ontology handed down through tradition obstructs new developments, especially in the basic modes of thinking. As long as the particularity of the conventional theoretical framework remains unquestioned, we will remain the toils of a static mode of thought which is inadequate to our present stage of historical and intellectual development.

Rifkin’s use of “the near-zero marginal cost” is a typical case of intellectual bad faith. His claim of capitalism’s inherent contradiction is the direct result of his Marxist confusion on economic calculation. The Marxist ideology⁸ assumes a static view of capital, as unproductive accumulation. He does not clearly announce his Marxist ontology. This is one feature of ideology: to proceed by successive concealments, not by clarifications. Accordingly, Rifkin’s confusion of the principles of cost calculation with the narrative of capitalism self-destruction conceals the operation of the new technology economy. In the development of new technology, capital is the object of sophisticated cost minimization, and possible increasing return to scale.

The result of the Wittgensteinian methodological detour is to clarify the costs of the production function and Rifkin’s ontology successively:

The following section 3 shows the changes in the marginal cost of the production of new technology can only be understood along with fixed and variable costs of production. This is clearly not what Rifkin is discussing.

3. Cost calculation in the production function

Rifkin uses the expression “near-zero marginal cost” 90 times across his book to address the costs of production and consumption of the sharing economy indistinctly. To clarify the meaning of this expression in economics, we need to pose the basic variables of the production cost model (VARIAN, 2020).

3.1. The firm’s fixed and variable costs

Let us start by defining the firm’s total cost as a sum of two variables: fixed cost (F) and variable cost C_v(y), assuming that Y is the firm’s output. The total cost of production can be written as follow:

The total cost of production measures these two costs per unit of output. The fixed cost measures the cost of the machines or the loan necessary to produce a product. In manufacturing or non-manufacturing, fixed costs may include the rent or the mortgage payment on building facilities. Here, we assume the fixed cost is calculated in the short run. It means that fixed costs tend to decrease as the output increases. In the long run, those fixed production factors can be considered “quasi-fixed” factors of production (VARIAN, 2014, p. 375). Hence, one would need to consider the wear of production materials, the change of factory size, or labor force size as increasing costs.

The variable costs are the changing costs of raw materials needed to build something. For example, those raw materials are subjected to market price fluctuations. The variable cost depends on the production output. If one produces one unit, the variable cost is 1. If it produces 2, the variable cost doubles. Hence, the average variable cost increases when the output increases.

Combining the fixed and variable costs, the total cost of a firm’s production produces a U-shaped average cost curve. If the production is more efficient, the producer may decrease the variable costs momentarily. However, variable costs will continue to increase sharply when the production scales up. If we consider both fixed and variable costs in relation to the firm’s output, we are also considering the marginal cost of production.

3.2. The marginal cost of production

Hal Varian (2014, p. 369) defines marginal cost as the change in costs divided by a change in output. The ratio of production, i.e., the marginal cost, is noted MC(y). This measure of productivity asks: how will the cost change given a change of output Δy?

Before presenting the marginal cost model, let us address Rifkin’s expressions: “zero marginal cost” and “near-zero marginal cost”. If the marginal cost is at zero, there is no production because the marginal cost is the sum of fixed and variable costs. In other words, to have zero marginal cost, the firm needs to be not producing anything (or zero units of production). Let us assume that Rifkin means that there is still some production, but digital production of one more unit brings the firm’s cost of production to a “near-zero marginal cost”. We will verify below if that is possible according to the marginal cost model of the firm. We assume here that the firm’s cost of production includes fixed and variable costs of production. The marginal cost is:

The marginal cost of production is a measure of a rate of change. Rifkin assumes MC is a static value given the cost of the output of one unit divided by the (indefinite) number of its users: close to zero when it comes to information technology. Since the marginal cost of technological output is a rate of change, it assumes an increase in output with a relative change in costs. So, the marginal cost is measuring a rate of change in production.

1- For the sake of Rifkin’s argument, let us assume the production of software when the cost of reproducing it is near zero, i.e., after the initial cost of writing the new code, and before any additional updates. At this precise moment, the variable cost of production of software is decreasing. It decreases because there is no additional “code writing material” involved in the production of one more unit of software. One software developer developing one unit of software is 1/1. To assume that the decrease of the variable cost of production of software is continuous, one must integrate that the same developer producing one unit of software distributes it through a digital network, let us say 1/10 000 users = 0,0001 - basically, when software is reproduced (or distributed) at almost no additional variable cost.

Even within this particular scenario of software production, there is no indication that the “near-zero marginal cost” of software production will bring the firm’s profit⁹ down to the point of bankruptcy, as Rifkin strongly advocates. The average variable cost cannot stay fixed because software production is one of the most dynamic fields of the technology industry.

2- The average variable cost will rise again due to the fixed factors of production, like updated computers, newly educated labor force, or extensive worldwide network of servers.

3- The average variable cost and the marginal cost are the same when producing the first unit of software, but they increase later as more software units are produced.

After this clarification of the cost function, one sees more clearly that Rifkin is not operating within the parameters of economic theory. Rifkin does not discuss any alternative cost functions. Section 4 below shows that he does a “selective reading” of marginal cost to set up his Marxist agenda.

4. Narrating deception: productivity as destruction

Since Rifkin is not engaging in calculus to demonstrate changes in costs in the production output of digital technology, his argument cannot be empirical. The only explanation is that he is working with scholarly references used essentially as instruments of persuasion.

To show that, we review all the references he uses in selected pages (RIFKIN, 2014, p. 10-15) to lead the reader to think that a near-zero marginal cost is sign of capitalism destruction (rather than a way to assess if the firm’s behavior is cost-minimizing). He coins the expression “extreme productivity” to dismiss the real economic definition of productivity. According to him, the “zero marginal cost of production” leads to a shrinking of the firm’s capital due to its capture by new technology communities. It implies that markets for corporations disappear for the benefit of the niche economies of technological communities.

We will reveal his uses of references in chapter 1, entitled The great paradigm shift from market capitalism to collaborative common (RIFKIN, 2014, p. 10-15). We take each element of his argumentation on marginal cost linearly and check all the references grounding its authority.

4.1. The dismissal of capitalism

To justify his argument on the dismissal of capitalism, Rifkin (2014, p. 10) uses Oskar Lange’s paper published in 1936 on “the economic theory of socialism”. In his selected extract (LANGE, 1936, p. 129-130), Lange talks about the difficulty of innovative firms recovering their R&D investment because of low returns on cheaper commodities. Lange tries to indicate that the condition of market competition, especially in innovation, is unsustainable. Innovative industries seek to restrict new entrants and establish monopolies to protect their investments. He suggests that the contradiction between long-term investment and value destruction through innovation is fatal to the capitalist system. As such, the state should be more efficient in employing innovative investment through public projects. For Rifkin (2014, p. 12), what matters is to maintain the idea of an inherent contradiction (schizophrenia) within the capitalist system, based on the nonsense that cheap commodities undermine productivity:

the inherent entrepreneurial dynamism of competitive markets that drives productivity up and marginal costs down. Economists have long understood that the most efficient economy is one in which consumers pay only for the marginal cost of the goods they purchase. But if consumers pay only for the marginal cost and those costs continue to race toward zero, businesses would not be able to ensure a return on their investment and sufficient profit to satisfy their shareholders.

Rifkin and Lange neglect the role of creative destruction (SCHUMPETER, 1942) as an alternative explanation of innovation. The unceasing change of firms by more productive ones is the rule in an open market economy. Further, Rifkin avoids the issue of socialist cost calculation, which is the topic of Lange’s paper. Lange argues that it is possible to calculate the investment rate, the volume and structure of public goods, and the price of all inputs without the market (VON MISES, 1990; KOCHANOWICS, 1990, p. 22-25; HAYEK, 2015; SCHUMPETER, 1942, p. 172-186). A central planning board determines those factors of production to minimize costs and make the marginal cost equal to price. Rifkin supposes that new technological communities play the role of Lange’s technocratic planning boards. This helps him to argue for a conflict between cooperative associations of techies and corporations. In today’s economy, corporations and tech communities coordinate productive activities requiring the same production factors¹⁰.

4.2. Technology unemployment

Rifkin (2014, p. 12) continues hammering his idea of the destruction of capitalism using Keynes’ (1930) concept of “technological unemployment”. Keynes argues that an increase in productivity reduces the cost of goods and services, but also the labor revenue to unemployment. No economist today believes the Luddites fables. As Brynjofsson and McAfee (2014, p. 176-7) pointed out, the reduction of working hours due to productivity has directly increased leisure time as well as the development of new forms of work (aka hybrid entrepreneurship, intrapreneurship (BRAUNERJHELM et al., 2018). The other issue is short-term maladjustment of skills and education, when firms are obliged to adapt rapidly to technical change. Here again, Rifkin does not address either labor flexibility (GRILICHES, 1997; HANUSHEK; KIMKO, 2000) or educational reforms (EKLUND; PETTERSSON, 2017; KRUEGER; LINDAHL, 2001) to make labor adjustments possible and avoid unproductive social costs.

4.3. Anomaly of the economic regime

Rifkin (2014, p. 12) cites Summers and Delong’s paper (2001) as a more recent source of analysis on the role of IT in the new economy. Those economists draw a similar picture than we have so far on the new technologies. Namely, they indicate that IT has pushed a new technological frontier in which growth effects are still unknown but important (as electricity was before, see (JOVANOVIC; ROUSSEAU, 2005). They indicate that new technology prices equal to the marginal costs. They argue that the condition to consider a marginal cost of distribution (we assume they mean the cost of reproduction) close to zero means that firms need to develop a new business model to find revenue sources. They suggest that new revenues models should not be supported by state allocations, but temporary monopolies. In theory¹¹, everyone knows that temporary monopolies permitted by technological advances allow firms to finance their R&D costs.

Rifkin sees a contraction in the need for a temporary monopoly within a competitive market. Every economist knows this is not the case. The dynamic of the production system accommodates different stages of technological development. Those stages are not fixed systems confronting each other. Summer & DeLong indicate they do not know the direction the IT innovation will take. Rifkin takes their comments to reveal anomalies in the viability of the economic “regime”. He crystallizes a time-sensitive moment of innovation as if it was a shift of a political regime. Assuming the political metaphor allows him to pound again his reader with his dream of the inevitable self-destruction of the capitalist system. Nothing in what Summers & DeLong said lead to that ideological twist. Entrepreneurs have found creative ways to deal with new products or services (with a low marginal cost of reproduction) by conceiving multisided platforms. Rochet and Tirole (2003) and Evans and Schmalensee (2016) have defined multisided platforms as a market where different types of actors (free users, marketeer, developers) can appropriately charge for the different services provided. Further research (ROCHET; TIROLE, 2003; 2004; 2006) has studied how business owners may charge a mix of membership and usage fees to cover their fixed and variable costs.

Concluding this section, our hermeneutic analysis reveals that Rifkin is suffering from a selection bias. For him, the anomalies of the capitalist system are so overwhelming that the entrepreneurs themselves cannot correct them. However, is Rifkin’s claim for a “paradigm shift” an adequate answer to the entrepreneur’s issue with costs?

Conclusion

In conclusion, we recall few of the issues of cost calculation Rifkin is so conveniently avoiding. We also ask an ethical question: what does it mean to ignore costs and substitute them with ideological shortcuts?

Rifkin’s treatment of “the near-zero marginal cost of production” selects some aspects of costs and disregards how to properly calculate them in economics. Besides the fact that Rifkin seems obsessed with advertising the ending of capitalism, there are other lessons to take from his renewed exercise in M arxist rhetoric today.

The first remark is an epistemological issue. The relation between new technology and economics demands more knowledge of the productive factors, not less. For example, suppose Rifkin talks of the “near-zero marginal cost” of new technology. In that case, he should start with basic principles: 1- the profit maximization, 2- the cost-minimizing, and 3- the return to scale of new technology products.

1- All firms or organizations follow the profit maximization principle by increasing their revenues above their costs. For that, they make sure that their cost of production (fixed, variable, and marginal costs) equates to at least its market price. 2- Further, the sharing economy is efficient by increasing revenues and simultaneously minimizing its costs. The IT industry has been driving new solutions to minimize production costs, by delocalizing software development, supplying complex systems, and developing open-source operating systems. 3- Finally, firms that decreased their average costs have generated an increasing return to scale. New tech firms have shaped the supply side by using specialized labor¹² to cope with combinatorial solutions in digital technologies. There are endless combinations, since the development of new technologies¹³ increases exponentially. The issue is and will always be to deal with the adjustments of the job market in relation to effective labor developing those technologies. Since the mid-1990s, the question of labor requires creative solutions to be able to process those new combinations (SIMON, 1971; WEITZMAN, 1998; BRYNJOLFSSON; MCAFEE, 2014, p. 82). Then, what does it mean to ignore costs?

To ignore costs means to give up economic and social responsibilities. Economic because when one removes the ability to perform production cost calculation, one removes the ability to evaluate the interaction between different factors of production. Social because the coordination of the factors of production also implies the best allocation of human resources among productive activities. Messing up with that equation is a very inconsiderate way to deal with human potential, i.e., the future productive capabilities of individuals.

References

ARROW, K. J. The economic implications of learning by doing. Review of Economic Studies, v. 29, p. 155-173, 1962.

BRAUNERHJELM, P.; DING, D.; THULIN, P. The knowledge spillover theory of intrapreneurship. Small Business Economics, v. 51, p. 1-30, 2018.

BRYNJOLFSSON, E,; McAFEE, A. The second machine age: Work, progress, and prosperity in a time of brilliant technologies. WW Norton & Company, 2014.

DELONG, J. B.; L. H. SUMMERS. The ‘new economy’: background, historical perspective, questions, and speculations. Economic Review-Federal Reserve Bank of Kansas City, v. 86, n. 4, 2001.

EKLUND, J. Varför engagerar sig inte fler ekonomer i debatten? Ekonomisk Debatt, v. 48, n. 3, 2020.

EKLUND, J,; L. PETTERSSON, L. Högskola i otakt. Dialogos Förlag, 2017.

EVANS, D.S.; SCHMALENSEE, R. Matchmakers: the new economics of multisided platform. Boston: Harvard Business Review Press, 2016.

GRILICHES, Z. Education, human capital, and growth: a personal perspective. Journal of Labor Economics, v. 15, n. 1, 1997.

HACKER, P. M. S. Wittgenstein’s place in twentieth century analytic philosophy. Oxford: Blackwell publisher, 1997.

HANUSHEK, E. A.,; KIMKO, D. Schooling, labor-force quality, and the growth of nations. American economic review, v. 90, n. 5, 2000.

HAYEK, F. A. The Road to Serfdom. London: Routledge Classics, 1944.

HAYEK, F. A. Collectivist Economic Planning. Auburn: Mises Institute, 2015.

JOVANOVIC, B.; ROUSSEAU, P. L. General purpose technologies. Amsterdam: Elsevier, 2005.

KEYNES, J M. Economic possibilities for our grandchildren. London: Palgrave Macmillan, 1930/2010.

KOCHANOWICZ, J. Introduction to the edition. Economic Calculation in the socialist commonwealth. Auburn: Mises Institute , 2015.

KRUEGER, A. B.,; M LINDAHL Education for growth: Why and for whom?. Journal of economic literature, v. 39, n. 4, 2001.

LANGE, O. On the economic theory of socialism: Part one. The review of economic studies, v. 4, n. 1, 1936.

MCGINN, M. Wittgenstein and the philosophical investigations, Routledge: New York, 1996.

MANNHEIM K. Ideology and Utopia, an introduction to the sociology of knowledge. New York: A Harvest Book Harcourt Brace and Company, 1936.

MARX, K. Capital. London: Penguin Classics, 1958/1976.

MISES VON, L. Epistemological problems of economics. Indianapolis: Liberty funds, 2013.

MISES VON, L. Economic Calculation in the Socialist Commonwealth. Auburn: Mises Institute , 1990/2012.

OGDEN, T. No Value. Stanford Social Innovation Review, Fall 2014. Available at : Available at : https://ssir.org/book_reviews/entry/no_value. Retrieved on: 12 oct 2021.

PASSMORE, J. A Hundred Years of Philosophy, London: Penguin Books, 1957.

RIFKIN, J. The zero marginal cost society: The internet of things, the collaborative commons, and the eclipse of capitalism. St. Martin’s Press, 2014.

ROCHET, J-C.; TIROLE, J. Platform competition in two-sided markets. Journal of the european economic association, v. 1, n. 4, 2003.

ROCHET, J-C.; J. TIROLE. Two-sided markets: an overview. Institut d’Economie Industrielle working paper, 2004.

ROCHET, J-C,; J TIROLE. Two-sided markets: a progress report. The RAND journal of economics, v. 37, n. 3, 2006.

RORTY, R. The linguistic turn: recent essays in philosophical method. Chicago: The University of Chicago Press, 1967.

RAYMOND, E. S. The cathedral and the bazaar. Cambridge: O’Reilly, 1999.

SIMON, H. A. Designing organizations for an information rich world. In: GREEN-BERGER, M. (ed.). Computers, Communications and the Public Interest, Johns Hopkins Press, 1971.

SCHUMPETER, J. A. Capitalism, Socialism and Democracy. New York: Basic Books, 1942.

VARIAN, HAL R. Intermediate microeconomics with calculus: a modern approach. New York: WW Norton & Company, 2020.

WEITZMAN, M. L. Recombinant growth. The Quarterly Journal of Economics, v. 113, n. 2, 1998.

WITTGENSTEIN, L. Philosophical grammar. Oxford: Basic Blackwell, 1974.

WITTGENSTEIN, L. Philosophical investigations. Englewood Cliffs: Prentice Hall, 1958

NOTAS

Author notes

E-mail: philippe.rouchy@bth.se