In the late nineteen fifties, Professor John von Neumann suggested that the human brain uses an internal language for mental calculation. He named that language the Primary Language. He suggested that the Primary Language differs from all the human (Secondary) languages used for communication. We consider Primary Science to be the science done with the Primary Language, that is differs from the familiar, conventional science.

The ultimate goal of this research is to reveal the nature of the Primary Language. To accomplish that, it is important to find those algorithms, used by the human brain, based directly on this language. The results of previous research revealed the existence of at least two ancient algorithms critical for the development of human intelligence. It is our assumption that in absence of the secondary languages, during times long past, those algorithms would have directly utilized the Primary Language. These two algorithms are Linguistic Geometry (LG) the algorithm for optimizing warfighting strategies, and the Algorithm of Discovery (AD) the algorithm for inventing new algorithms.

The main hypothesis of this research is that the Primary Language is the "language" of visual streams (mental "movies"). Another hypothesis states that the AD is a universal algorithm used for making discoveries. It suggests that the AD is based on multiple thought experiments, which manifest themselves via visual streams. It appears that visual streams are the only interface to the AD. The AD operates with three classes of visual streams: observation, construction, and validation. These visual streams can run concurrently and exchange information. Each stream may initiate additional thought experiments, program iv them, and then execute them. Visual streams are used by the AD to construct new algorithms and, in this way, make discoveries.

The current objective is to further investigate the AD by applying it to various discoveries, from different domains, such as Computer Science, Molecular Biology, etc., in order to reveal its inherent details. During this research, it was found that some components of the AD are utilized for every discovery, while others may be utilized for specific discoveries only. Once the major elements of the AD are revealed, generalizing them will lead to a complete understanding of the AD.

A comprehensive understanding of the AD and its components will lead to the implementation of the generalized AD, i.e., the final goal of making discoveries on demand. The expectation is that the implemented AD will have a profound impact on all branches of science, including Computer Science, and, in particular, Artificial Intelligence (AI).