PCA is a special case of factor analysis, so you are representing them as observations of a latent variable (which is often a narrative people use when explaining why two x variables are correlated)