One of the most common features of disk galaxies in the local universe is the presence of prominent spiral arms. Despite major developments in the 1960s and decades of studies, we still don’t know what the necessary and sufficient conditions are for spiral arm formation. Current efforts on the theory of spiral arm formation primarily focus on testing variations of density wave or swing amplifications on nearby galaxies. Progress in this field is slow. Breakthroughs can come from observations of high-redshift galaxies when spiral arms are in the early stage of formation. In this talk, I will present some of my first results on integral field spectroscopic observations of the most ancient spiral galaxies at z>2, including gravitationally lensed galaxies. The high-spatial resolution kinematic fields imply a population of rare spiral galaxies living in dynamically cold environments at high redshift. I will discuss how ongoing follow-up IFS observations on a larger sample of spiral galaxies at z>2 can help to constrain the earliest onset mechanisms for spiral arms.