A matter of great debate, arguments on both sides, both convincing enough, but you have to make your decision when you read about each individual case. Good descriptions exist from examples all around the world. Some of the authors, Kerrich, Groves, Robert, Kretschmar.
For centuries, tens of thousands of geoscientists are working on deciphering the origin of each individual gold deposit. Hundreds of thousands of publications report on the difficulties to find easy classifications. The bottom line is that you can find no complete match between two gold deposits.
Therefore, when it comes to the genesis of gold deposits, a simple "either" - "or" question can not be answered appropriately in a few sentences.
you raise a "chicken-and-egg question". But it may be answered in a simple way: Mineral deposits are geological bodies, including orogenic gold deposits. Therefore you need accommodation space to emplace these gold deposits. Following the logic, the presence of shear zones is a prerequisite for the gold deposits to be emplaced. David I. Groves has demonstrated this in a very good way and presented this together with other geologists dealing with this subject matter very well (see the first answer).
while it is true that orogenic gold mineralizations (not only the Archean ones) are controlled by shear zones, it is also true that not all shear zones of orogenic belts are gold-bearing.
A critical point for the location of the gold-depositing site in orogenic belts is where the Au-transporting fluid is generated at depth. That factor controls which set of shear zones (at regional and local scales) is ultimately going to host the Au-fluid, and therefore generating the deposit. In the geological literature - e.g., that of the European Alps - there are beautiful examples of sets of shear zones that are identical to those found in orogenic deposits; however, they are totally barren. Clearly, this is evidence that the structural conditions for generating conductive shear zones are met in several regions of orogenic belts; yet, those conditions are not sufficient to form a deposit.
The answer to your first question is that orogenic gold-bearing lodes are definitely controlled by the shear zones and not the vice-versa.
Your second question is an interesting one; and I would like to elaborate a bit. Orogenic gold lodes are essentially controlled by shear zones. However, lithology has an indirect control over deciding the location of ore precipitation at different levels of crust. For example, it has often been noted that the boundary-zones between rock types of contrasting rheological properties are more susceptible to fault/shear development, through which eventually ore fluids propagate. So, such cases infer that even if the ore-bearing lodes are shear-zone controlled, lithology did play an indirect role by providing a favorable rheological environment to develop shear zones. Also, it is a common observation that in gold-endowed granite-greenstone belts, ore lodes are mostly hosted within greenstone rocks rather than granitoids. In fact, this spatial isolation between granitoids and gold-bearing veins led many to reject any genetic relationship between them ('The granite-gold connection ?') . However, it could be just an expression of rheological contrast between granitoids and greenstones during regional deformation, where the latter preferably tends to develop weak (shear/fault) zones, which eventually guide the passage of ore fluids and formation of gold-quartz lodes. Also, mafic and ultramafic rocks have better potential to enhance alteration processes (e.g., carbonate replacement alteration) eventually leading to deposition of gold, which may be due to relatively higher concentration of cations (Fe, Mn, Mg, Ca) in them as against granitic rocks (Kerrich and Fyfe, 1981). Hence, deformation-induced shearing is the main controlling factor and lithology has just an indirect control over deciding the location of ore lodes.