Abstract: Swarm antenna arrays, composed of spatially distributed antennas mounted on unmanned agents, offer substantial flexibility for wireless sensing and communication. However, their reconfigurable architecture, susceptibility to collisions, and inherently stochastic nature present significant challenges to achieving collaborative gain. This paper investigates the feasibility of achieving coherent beamforming in such systems from both deterministic and stochastic perspectives. First, a rigorous theoretical framework is developed to characterize the necessary and sufficient conditions for the emergence of grating lobes in multiple linear configurations and further show that for dual linear arrays, the classical half-wavelength spacing constraint can be safely relaxed without introducing spatial aliasing. Second, a theoretical analysis, supported by empirical validation, is presented and demonstrated that coherent gain can be approximately preserved under realistic positional perturbations. The proposed results reveal that spatial perturbations introduce measurable degradation in the main lobe, an effect that cannot be mitigated merely by increasing the number of antennas. Instead, the primary benefit of scaling lies in reducing the variance of perturbation-induced fluctuations. Finally, the emergent deterministic behavior of large-scale disordered arrays is examined by analyzing the spectral properties of the associated Euclidean random matrices. These results provide new theoretical foundations and practical design guidelines for enabling coherent functionality in swarm antenna arrays.