A better understanding of the fluid dynamics of disease transmission by disintegrated respiratory droplets has been the focus of great attention since the recent outbreak of COVID-19. In particular, human respiratory activities such as coughing, sneezing and even talking and eating expel a large amount of pathogen-laden droplets. Particularly, during eating or drinking, the physical properties of saliva can be changed. In this study, we investigate the atomization morphology of expelled artificial saliva mixtures from the perspective of varying fluid physical properties, specifically surface tension and dynamic viscosity. Using high-speed shadowgraph experiments on artificial saliva, we visualize and analyse the disintegration of saliva liquid sheets into ligaments and droplets. We find that the viscosity and surface tension affect the droplet size formed from expelled saliva and follow scaling laws that have been previously observed and predicted for constant shear viscosity. We conclude that the changes in physical properties of saliva induced by eating and drinking tend to favour the formation of smaller droplets during sneezing or coughing, which could drive the airborne transmission pathway of pathogens. Furthermore, we derive a theoretical model based on scaling arguments that shows the breakup time of ligaments produced from the artificial saliva mixtures is dependent on the capillary number.