You are raising a serious question. The classical school concerned with value considered production of commodities by means of commodities. A modern leader of this school was Piero Sraffa, who wrote a book in that tradition of economics called “The Production of Commodities by Means of Commodities,” Cambridge University Press, 1960. If we think about agricultural production, it is easy to see that it take seed-corn as input to produce corn as output. As Paul Samuelson puts it “the goods that can trade are also by definition inputs that can trade.” Samuelson, 2011, Collected Paper, V.7, p. 436.
I recommend that you check to see if your model fits within such classical line of thought to strengthen the design of your model. Good luck in your research.
I would say yes, it is possible. in fact, we have some dual-role criteria which can be considered both as an input or output. Fore instance, in Network or Dynamic DEA) I can refer to semi-finished products which after exiting from period P (as an output) immediately enters to period P+1 (as an input). Green R&D cost is another example which generally would be an input factor. But, if such a cost tangibly leads to an innovation or contribution in production process (or other departments or aspects of products or services) in next periods, so that we see some "competitive advantages" such as a reduction in the production costs or an increase in the quality of products and services or even an environmental achievement; these can be considered as output in the subsequent periods.
I cannot suggest this idea. Although this specification is very involve with input output uncertainty, but still understandable. I would therefore suggest to read:Article Data Envelopment Analysis with Uncertain Inputs and Outputs
It is not logic at all that the same variable used to be both input and output in the same time, means author don't know the object of the model /study.
In the production system inputs are different to outputs in all sector, i think. If the case that you study the inputs are same outputs, this is other thing you need to explain.
In addition, in the production function there is transformation of input to output, if used the same means there is no production, which is not the case in reality. Also, in the case of cost function is not logic.
in my opinion it can cause the problem of heterogeneity, so i may suggest you to use different variables for inputs and outputs. or use use the proxy variable in one case.