We present an overview of the role, benefits, and drawbacks of theory in scientific research, particularly in the social and behavioral sciences. We discuss what theory is and what it is not. We also focus on some key elements of theory such as its ability to explain phenomena at multiple parallel levels of analysis. Evolutionary theory is offered as an example that illustrates the importance of conceptual integration across different disciplines. We further describe the key characteristics of good theories, such as parsimony, depth, breadth, and coherence (both internal and external), and we encourage the use of “coherence stress-tests” to help refine theory. We then discuss 4 advantages and 10 disadvantages of using theory in social and behavioral science research. Finally, we suggest conceptual tools and provide a list of recommendations for theory-driven research. We hope this chapter will help in the complex pursuit of improving research practices in the social and behavioral sciences.

All aspects of linguistic knowledge are ultimately based on speakers’ experience with lexical expressions, but of course, knowledge of language, notably, grammar, exceeds their memory of particular lexical tokens. It is a standard assumption of the usage-based approach that grammar involves a taxonomic network of constructions ranging from prefabricated strings of lexical expressions to highly abstract schemas. Chapter 4 describes the taxonomic organization of constructions and their development in L1 acquisition and language change. It includes a detailed discussion of current research on statistical grammar learning in infancy, the acquisition of constructions during the preschool years and two case studies on the rise of constructional schemas in language history.

Chapter 6 is concerned with symbolic associations between form and meaning. Cognitive linguists have analyzed the conceptual foundations of linguistic symbols in great detail, but they usually look at the semantic pole of lexemes and constructions from a synchronic perspective. In the dynamic network approach, the focus of analysis is on the development of symbolic associations. Combining frame semantics with research on symbol learning in L1 acquisition, Chapter 6 argues that symbolic associations evolve from recurrent paths of interpretation that become entrenched in memory as a consequence of automatization. The proposed analysis challenges the traditional distinction between encoding and inference and provides the basis for a dynamic theory of meaning in which linguistic elements are seen as cues or stimuli that activate a specific concept, i.e., the figure node, of an open-ended network of world knowledge. Since the activation status of the network varies with speakers’ (and listeners’) mental states in a particular situation, linguistic cues can give rise to different interpretations in different contexts.