This study aims to provide an extended interpretation of Chomsky``s (2013) Labeling Algorithm. Chomsky(2013) assumes a fixed labeling algorithm that licences syntactic objects. It is generally considered that the label can be one or the other of α, β, which is an asymmetric label. I propose that Labeling Algorithm should involve a symmetric label. That is to say, both α and β can contribute equally to labeling. Symmetric label can be determined by the intersection of α and β through Agree (i.e., Φ-features) and the union of α and β without Agree (i.e., categorial features of α and β). The second option is a way to determine the label that does not contradictory to α and β. Concretely, Labeling Algorithm can determine the label projecting both α and β, when α and β have in common the same categorial features without Agree between them. Furthermore, I claim that symmetric label is useful for an analysis of serial verb construction and Spanish secondary predicate construction. I provide the extraction phenomena out of α and β to show the fact that symmetric label makes the elements of α and β accessible to further operations.