The Construction of the Past: Towards a Theory for Knowing the Past

information Article The Construction of the Past: Towards a Theory for Knowing the Past Kenneth Thibodeau Tim Tam Trail, Evergreen, CO 80439, USA; Received: 2 October 2019; Accepted: 23 October 2019; Published: 28 October 2019



 Abstract: This paper presents Constructed Past Theory, an epistemological theory about how we come to know things that happened or existed in the past. The theory is expounded both in text and in a formal model comprising UML class diagrams. The ideas presented here have been developed in a half century of experience as a practitioner in the management of information and automated systems in the US government and as a researcher in several collaborations, notably the four international and multidisciplinary InterPARES projects. This work is part of a broader initiative, providing a conceptual framework for reformulating the concepts and theories of archival science in order to enable a new discipline whose assertions are empirically and, wherever possible, quantitatively testable. The new discipline, called archival engineering, is intended to provide an appropriate, coherent foundation for the development of systems and applications for managing, preserving and providing access to digital information, development which is necessitated by the exponential growth and explosive diversification of data recorded in digital form and the use of digital data in an ever increasing variety of domains. Both the text and model are an initial exposition of the theory that both requires and invites further development. Keywords: archival bond; archival science; class diagram; constructed past theory; intentional domain; pragmatic information theory; sphere of interest 1. Introduction The past does not exist and never did. The past is always something that is constructed by thinking, writing or speaking about former times. In this respect, knowledge about the past is not different from knowledge gained from direct sensory experience. Neuroscience tells us that the reality we perceive is not a direct reflection of the external world. Rather, perceptions are constructions produced by the brain’s predictions or guesses about the causes of incoming sensory signals [1]. The fundamental difference between knowledge gained from perception and knowledge of the past is not in the processes that produce them, but in their sources. Knowledge of the past comes from vestiges and imprints, where vestiges are persistent objects that survive from former times and imprints are previously generated constructions of the past. ‘Constructing the past’ may bring to mind the writing of history or biography, but the process can take many different forms: management review, audit, courtroom argument, archaeological analysis, psychotherapy, and others. One might assume that all constructions of the past, like these examples, are retrospective; however, even real-time activities, such as keeping a diary, video surveillance, medical registries, live news coverage, and web crawling, produce vestiges of the past by collecting and keeping certain types of data. Constructions of the past are necessarily incomplete and slanted. Consider the real-time activities just mentioned. Diary entries are selective. Video cameras have limited scope and finite precision. Medical registries are self-selecting. Live news coverage is typically short and limited to topics of presumed public interest. Web crawlers capture data about only a fraction of websites and do Information 2019, 10, 332; doi:10.3390/info10110332 www.mdpi.com/journal/information Information 2019, 10, 332 2 of 22 not follow all hyperlinks. Such limitations only increase with the flow of time as the vestiges of past times erode and disappear and imprints more readily bear the perspectives and biases of the contexts in which they are constructed. In fact, both retrospective and real-time constructions of the past are inevitably colored by the imposition of expectations and present purpose. They are often influenced by the results of prior constructions. Furthermore, they cover only limited time periods; address particular interests; use specific sources and finite amounts of data; and apply distinct methods for organizing, processing and rendering data about the past. This paper proposes a model of the development of knowledge of the past, called Constructed Past Theory (CPT). The domain of CPT encompasses only the process and materials of construction, not the knowledge that results. CPT does rest on certain assumptions about knowledge, most fundamentally the semiotic proposition that to know is to affirm a proposition, where a proposition is an expression that has an objective meaning that can be believed, doubted, denied or asserted as either true or false. It follows that knowing is a process, not a thing. The persistent counterpart of knowing, what we call knowledge, is either a persistent representation of one or more assertions, such as in a document, or the capability for producing or reproducing assertions. This capability assumes (1) a store or stores of data structured in a way that enables their retrieval and presentation in propositions, and (2) the ability to select, retrieve and process data in order to output propositions. CPT focuses on processes that construct the past from vestiges left behind, rather than those that collect or record real-time data. Its objective is to develop a framework for the discovery and delivery of vestiges from the past, the evaluation of their appropriateness for the purpose a construction has, and their exploitation in the process of construction. CPT also offers the potential to reveal and offset inappropriate preconceptions or predispositions by illuminating the original contexts in which things happened or existed in the past. Although a basic goal in developing Constructed Past Theory is to support implementation in automated systems, the theory does not accept the distinction between data and information that is common in information technology, as reflected in assertions such as: “data has no meaning or value because it is without context and interpretation;” and “data are discrete, objective facts or observations, which are unorganized and unprocessed, and do not convey any specific meaning.” Such assertions are contrasted to information, which is described as “data that have been processed so that they are meaningful;” or “data that have been interpreted and understood by the recipient.” [2]. Data are structured or, more accurately, exist within structures that enable their interpretation. Without a specified interpretative structure, it would not be possible consistently and coherently to perform even the basic database operations of create, read, update and delete. “1734” and “Peter” are marks, not data. ‘Datum’ is the Latin singular of data. It means a given. The answer to the question of what these two signs are can only be that they are strings of numbe (...truncated)