The purpose of this panel is to explore further our understanding of the ?work? entity and its role in information retrieval. A work, at a basic level, is a deliberately created knowledge-record (i.e. a text, and oeuvre, etc.) representing a coordinated set of ideas (i.e., ideational content) that is conveyed with the purpose of being communicated to a consumer. Research into the nature of works has begun to yield both empirical and theoretical understanding of the nature of the ?work? entity. The importance of the work to IFLA?s FRBR entity-relationship model represents a major milestone in the history of bibliographic retrieval. Other metadata models, such as the CIDOC CRM, seek to represent the work notion in the context of systems for organization of cultural heritage artifacts and their representations. In fact, the concept of the work has been critical to many disciplines over time. These notions of the nature of a work are complementary and help provide important contextual information for the design of work-centered information retrieval. In 2003, a panel on ?works as entities for information retrieval? presented basic definitions and demonstrated operational concepts of ?works.? The panel proposed here was developed from the question and answer discussion period following that panel. In this second look at the work entity, we will visit interdisciplinary concepts of the work, which demonstrate some of the ways in which works help bridge cultural boundaries. Richard Smiraglia will describe the concept of the musical work from semiological and epistemological perspectives. Jonathan Furner will describe the concept of the work in the visual arts, exploring developments in the ontology of art and cultural informatics. Birger Hjørland and Jack Anderson will bring speech act theory, social semiotics and related perspectives to bear on the distinction between concepts of ?work,? ?document,? and ?genre.?