Summary :
This project will produce a database as a hypertext file assembly (a collection of files with related contents of various types). It will develop general tools and techniques for making sense of the interrelation of the contents in those various files. The first embodiment of the database will be in Macintosh Hypercards. The data of the files will be encoded for platform independence to permit portability to client- server facilities for the second embodiment. The scheme used will be Hyper Text MarkUp Language for compatibility with the World Wide Web. Both Macintosh-based and network based versions of the database will be produced during the period of the project. These will be published as machine readable files with the software needed for their use in the different environments. Prototypes of this file structure have been developed in past years. (See appendix A.)
The database to be produced includes an extended case study of cognitive development in the computing environment. The case study corpus, the text of interpretations, graphics, video, audio, and computer based models are owned by the principal investigator. The interpretations of the case study and related articles were written by the principal investigator. (See appendix B for more detail.)
Project Description - (Sections 1 through 4)
1. A Question of Appropriate Scale
1.1. There is increasing interest within the government and among the public for broadening access to on-line information.[1] Yankelovich (1992) argues correctly that "scaling up" is an essential concern:
"...although today's electronic materials exist primarily on a small scale -- book size -- tomorrow's collections will be much larger -- library size. One fundamental challenge for the next five years, therefore, is to ensure that the electronic book technology can scale up...."To the librarian, books look small. To the scholar, even individual books may loom very large indeed. Between the book and the library are intermediate scales of significant size, such as the scholar's collected works or the collection of information related to a specific issue. Granted that any on-line tools developed should have broad generality, one needs ask how the tools would be used, by whom, and for what sorts of purposes.
1.2 Making Connections and Making Sense
Electronic mail, the easy dissemination of large bodies of text from FTP-accessible archives, the capability for searching large collections of material achieved by the WAIS project, and the introduction of the point-and-click GOPHER interface inspire the hope that we are on the verge of breakthrough in use of on-line information over electronic highways.[2] Such is plausible, to the extent that efficient search and making connections among distributed items is central. However, after "making connections," scholarship requires one to make sense as well. As we scale up from the on-line book to larger collections, we need pay attention to ways of supporting the sense-making goals of scholarship. If we work with intermediate sized collections of on-line files, we can explore the problems of making sense without being immediately overwhelmed with the difficulties of filtering out the flood of irrelevant possible relations.[3]
2. A Hypertext File Assembly
2.1. I propose to develop a hypertext database.[4] The database will be a file assembly, that is, a collection of separate but closely related files. The methods and tools to be developed in this project are focused on user-constructible cross-file-linking journals. (These are described in detail in following sections.)
2.2. The first database embodiment will be in Macintosh Hypercards, because that is a widely available and easily mastered software environment. We will insulate the database against vendor dependence, software environment dependence, and medium-specific dependence. The text portion of the database will be able to be recreated at will by loading from ascii text files. Graphics parts will use standard formats. Looking beyond the initial implementation, the text markup and the specification of structural relationships within and among different files and their elements will be encoded for compatibility with the World Wide Web, that is according to SGML (standard generalized markup language) and its extension HTML (Hypertext Markup Language).[5] This encoding scheme will permit transfer among different platforms by the writing of implementation-specific load and download programs. Subsequently, when the database is installed on a network server, it will be made accessible for workstations by developing a client specific interface.
2.3. The project will produce an example of a mid-scale database employing such a coding scheme. At the end of the project it will be portable between a version based on local storage media and a networked version. The local version will be published either as primarily text-files with some graphics and programs on a 20 megabyte floppy or as a CD-ROM (with additional graphical materials). The networked version will be published and accessible via a national network. This database will be created with material owned by the principal investigator. Software needed for its use will be distributed with the database.
3. Details of the Database to be Developed
3.1 The Materials of the Database
3.1.1. The materials can be considered to be of five types: text, graphics, audio, video, and programs. [6] Each type is important, but text is primary. The textual materials are comprised of an extensive case study corpus and a suite of interpretations of those corpus materials. The focus of the case study is the impact of computing experience on cognitive development in two young children. The studies were undertaken at the MIT Logo Laboratory in the years between 1975 and 1981.[7] Because personal computers were not generally available then, the materials are complete with respect to the children's experience of computing. Interpretations, primarily textual material, have been published as books and articles in various places in the ensuing years.[8]
3.1.2. The programmed materials are of two sorts. One collection of programs are data from the study, typically Logo procedures used and/or composed by the subjects. The second set of programs are Lisp encoded models of ideas developed through interpretation of the corpus. The graphical materials are either drawings made by the subjects (both with and without computers) or photographs (of the subjects, of their surroundings, and of computer generated images). The audio/video component includes an extensive collection of materials recorded in one six months period, both in the laboratory of the Logo project and outside of it; some of the recordings are experiments and others are working sessions or play with the computer. Some of the original 70 half-hour videotapes have deteriorated, but most are still usable now.
Figure 1 - Relations Among Types of Files
3.2.1. A first view of the database is that it is a pyramid[9] (See Figure 1.) The broad base is data -- observations made on the collected materials of the case study corpus. The observational data are extensive and various. The next layer is comprised of interpretations of data developed from the corpus. These texts include observations of behavior, explanations of contexts, and discussions of meanings; the disciplines to which the study most closely relates are Psychology, Education, and Artificial Intelligence. These are collected writings of the principal investigator derived from and related to the specific case study corpus. The apex of the pyramid is the set of programmed models of ideas developed in the interpretations. This modeling layer also includes presented papers and articles in which the models are described. A portion of the project effort will be focused on extending earlier models and creating alternative models for the interpretations. Within this pyramid, some materials are related "horizontally." For example, arguments of different articles in the interpretations relate very directly to one another, regardless of their publication in different books or journals. Other materials are related "vertically," as interpretations depend on the observations they purport to explain. Similarly, models depend on the interpretations on which they are based. These relations will be specified by hypertext cross-file links (see knowledge-links below).
3.2.2 The Journal: A novel feature of this database structure is the central role assigned to a journal.[10] Structurally, the journal is central to the formation of cross-file links. It is the place where the central nodes and pointers of the knowledge-links are stored. It is a structuring tool for use by the user. The content of the journal entries is composed by the user. In addition to cross-linking pointers, the journal may contain any other information the user wants to put there.
3.3 Searching in the Database
3.3.1. Browsing and skimming are normal modes of search with text and hypertext alike and will be possible. Within individual files of the database, index links will relate segments of text, as by a common thread running through chunks of the text. These index links are equivalent to terms found in the index of a book.[11] Functions permit sequential search for the chunks of text by following the index links between individual cards of the file until the index is exhausted. One difference between a book index and those proposed for this database is that each "card" has its own index of locally relevant index terms. This permits index use from any card of the file.[12] Index terms may have new references to other cards added. New terms also may be added to the index. Such added terms and references are integrated into the indexing facility by compilation whenever desirable. When multiple files are open in different windows, index links may be followed in different windows concurrently. Thus searching on the same or similar index entries in multiple files will permit comparison and contrast of chunks of text similarly indexed.
3.3.2. Knowledge Links: An additional form of link will be a feature of this database. It will be especially valuable in database uses involving comparative analysis of multiple texts. This link is named a "knowledge-link" of K-link.[13] The primary node of this link
Figure 2 - The Structure of a Knowledge-Link
is a text string composed by the user specifying what he sees as significant in the comparison of the other elements linked together. This text string is given a label, which is used as a key for accessing the linked nodes. The remaining nodes of the K-link are scoped elements of any sort within the other open files of the database. (See Figure 2.) These nodes are identically labeled. As used in search, specification of the K-link's text key brings into view in each open window the elements contrasted when originally linked. Maintenance of a text-key index structure keeps access times tolerable. (The creation of these links is discussed further below; see section 3.6.3.)
3.4 Analytical Uses the Database
3.4.1. It would be possible and useful to analyze a single file with the special facilities of this database (index links, the journal, and knowledge-links). In this circumstance, the index links would be the most useful feature because they would permit the user to specify new index terms which are logically equivalent to typed links. As the user experiences his/her own insights while reading and thinking about material in the file assembly, terms can be inserted into the local index; later the individual file index can be recompiled at an appropriate time. The journal facility would function primarily as the user's notes on the meaning of the elements, linked to the referenced chunk of element.
3.4.2. The application to which the journal and knowledge-links are primarily directed is multi-window/multi-file comparison. The K-links provide cross-file inter- connection; the user composes his specification of what is significant in the relationships among the particular elements linked within the file assembly. In general, the intention is that when links are formed, the links will point along path names into the character strings of other files. As the contents of such journal files become extensive, performance will degrade and file reorganization will be necessary. The format and content of information saved during link creation and its reorganization is discussed below (see section 3.6.4.).
3.5 Constructive Uses of the Database
3.5.1. The interpretation of case study corpora involves the exploration of empirical material to produce an interpretation, i.e. it involves theory construction. The theory under development at any particular time may be overturned or may require modification as additional materials are examined. Case study analysis would benefit from development of a system to support an iterative process of theory guided exploration, theory modification, and further exploration with the guidance of the modified theory. (see Figure 3.)
- Figure 3 - A Theory Development Cycle [14]
Such a facility should preserve the interconnection of interpretations and the particular grounds of those interpretations; this is the provision of an audit-trail of reasons for the current state of the interpretations. Further, if functional models are constructed based on those interpretations, aspects of those models should be linked also to the particular grounds for assigning each element of the model its specific character. This will permit a system for the progressive construction of computational theories open to modification under the regimen of regression testing. When a theory, viewed as a collection of structures and functions is changed for some new good reason, a process can be developed to verify that the theory still applies over all the cases to which it should apply. This will guarantee that no regression of coverage of particular cases has occurred.
3.5.2. Shared Data with Various Theories: With the database and facilities to be developed in this project, it is quite conceivable that different interpretations for the same corpus materials could co-exist within the database. The simplest case would be one wherein an early interpretation was shown to be flawed, but can only be partially replaced by a successor interpretation. Theories change over time, even those of one individual. Even more it is the case that other people -- if interested in the same corpus materials -- might develop an alternative interpretation of corpus materials that would be radically different from those of the corpus creator. Such are situations of shared data supporting alternative theories. Conversely, theoreticians relish the opportunity to extend the reach of applicability of their theories. One can imagine and certainly even hope for situations where common theoretical notions can be applied as interpretations to disparate case study corpora. Both of these circumstances would place a more rigorous requirement on theorists and analysts to articulate all the details of their interpretations and on corpus creators to share all the materials at their disposal.
3.6. Journal Files and Their Reorganization
3.6.1. The journal files described here are meant to recreate some of the value of journals in the computing medium but at the same time exploit the unique features of computing to provide support for enhancement of analysis and theory development. Journal files are intermediate products, "means," as they always have been, not "ends" themselves. This implies there will be a process through which the content of journal files will be examined and reorganized into some other sort of thing. As an analogy, the K-links of this project could be likened to short term memories; eventually, some are lost or discarded and some are converted into longer term memories with a different level of integration with what else is known. Happily the analogy fails. Human short term memories are lost forever; journal-stored knowledge-links can be kept as long as they may have some utility, so long as resources permit.
3.6.2. Knowledge-links and preserved context: The current design of this database facility assumes that any particular knowledge link will be created within an assembly of open files. The advantage of forming links into the particular assembly of files is that access to the original element permits an enriched appreciation of the context of the elements of the nodes. When some portion of a journal file is being reorganized, it is expected that the information in the knowledge-links will be extracted from the assembly of files (and thus abstracted from their context) and be preserved in a different form in the reorganized, secondary file. Consider the following as a specific example of a K-link in a journal file and its transformation under reorganization into a secondary file.
3.6.3. A concrete example of a K-link: In chapter one of Computer Experience and Cognitive Development, Lawler proposed a form of "cognitive motivation" as driving the activation of mental schemes. Specifically, he states that when a scheme is activated but fails to function as expected, the scheme becomes a "demon procedure" ready to become active again in any context sufficiently close to that of the original activation to provide a match above threshold at some later time. Let that general description represent the text at one node of the knowledge-link. In the corpus supporting the interpretation, one can find pictures of computer drawings made during the case study and programs made by one subject, a boy of eight. These materials could serve as another node of the K-link. A third node of the K-link could include selected functions in the models of the interpretations, specifically indicating the scheme activation routines. The central node of the K-link would be composed by the user in the journal file - perhaps with a text such as this: "The author's theory of scheme activation is based on a series of particular choices made by the subject in a drawing project; the example is adequate to make the idea clear but certainly is not coercive. If however, one could find other corroborating examples in this corpus or cite other well-developed psychological theories advancing this view, the argument should be taken seriously. It might well be compared with the activation methods advanced by Newell et al. in SOAR." After creation of this K-link text, the user would specify a text-string as a label. It would then be inserted in index sub files of each of the separate files in the assembly, permitting rapid access to the elements in all the files at any later time through its use as an access key. [15] Figure 4 graphically exhibits such a link in a file assembly.
Figure 4 - A Knowledge Link in a File Assembly
3.6.4. A K-link after reorganization: Reorganization will be undertaken by the database user. First the journal files will be saved in a archive. Next the K-links will be reorganized for inclusion in a canonical hypertext database, such as the Xerox' Notecards system or Civilized Software's Storyspace. Figure 5 shows the records of Figure 4 recreated in a secondary file. In the new file structure, copies of the original materials are pasted on cards and they are inserted in fileboxes. One card will have the picture and Logo code. Another will have some chunks of Lisp code from the model. A third will contain Lawler's description of his notions. Typed links are inserted between chunks formerly related as a collection. The links retain and further specify relations between the elements in the new structure. A card, logically the central structure and derived from the journal entry, will contain the users comments as well as date and time stamps, path names to the original files and pointers to the strings, and the original key-string made by the user. In short, the selected elements will be included in the secondary file, but the contexts of the original elements will now be in files that may no longer be accessible. Indicative information will be included With the central card permitting the reconstruction of the K-link if that should be required at some later time.
3.6.5. Journal-based Knowledge-Links as Link-Hubs
Hann et al. specified that the Link Hub should serve four primary functions. First, keeping a list of pending links. Second, Managing the creation of links. Third, knowing how to follow a link. Fourth, providing link information to a link-browser equivalent to the Intermedia Web View. The first three functions exist in prototype form. Following this line of development is a straight-forward way to begin making such facilities accessible and comprehensible. Working with concrete examples of the conversion of information from journal-based K-links into a canonical-form hypertext will provide information and insights into how well the process can be automated.
Figure 5 - Knowledge Links Reorganized
Materials from the original files have been abstracted into Notecards, filed within boxes, and connected with typed links as part of the reorganization.
4. Impact on Scientific Research and Education
The existence of such facilities as described here could help move case study analysis from its current place as a would-be science across the frontier into scientific credibility. The essence of that agenda is to provide interpretations of specific cases to support the creation and illumination of general laws. The agenda is an old one, and reasons for pursuing it were eloquently argued by Kurt Lewin (1935). It also was a recurrent theme of Artificial Intelligence (AI) in the late 1970's. [16]
Members of the AI community should be among the first to recognize the value of such a database facility for grounding details of computational theories in common experiences and issues of concern. Social scientists will recognize the value to supporting their arguments through having available exquisitely detailed examples accessible through unobtrusive hypertext links. Psychologists will recognize the value of having a "workbench" for their advanced students, where the students will be able to compare interpretations published in the literature with detailed case study corpus materials. These students will have the opportunity to second-guess the interpretations of major theorists, based on the data originally available. Finally, given the current interest in qualitative studies for educational research, the demonstration through such detailed databases of what is required for a thorough-going examination of the relationship between concepts and contexts may give concrete substance to the discussion of situated cognition.
The project work will begin in the academic year1993-94, the sabbatical year of the Principal Investigator. Following the sabbatical year, the PI will work on the project for three summers, in a period of 27 months. The staff of the project will be engaged with it for the last 24 of those months. In thios specific sense, this is a two year project. The work of the project will involve five categories of activities. They are tool development, database creation, content analysis, model construction, and distribution. Table 1 displays the expected schedule of effort of the PI and colleagues for the two funded years of the project. The PI will undertake content analysis of the material and will design the database, the tools, and the models created from the analyses. He will define, guide and oversee the work of the graduate student programmer (half-time) and the two undergraduate students (quarter-time). He will be responsible for design, production, and dissemination of the research products of the project. During the spring of 1994, when the PI will be on sabbatical at the MIT AI laboratory, he will determine which specific workstation platform will be optimum for use in developing the client-server version of the materials.
4.2 Anticipated Products
In each case mentioned below, the tools and programs necessary to use the materials will be published with the database content. The first product will be an on-line hypertext assembly made available in Macintosh Hypercards. A prototype version of this product is planned for September 1994. After use and evaluation by graduate students in educational computing at Purdue, an improved version will be published by September 1995. The Macintosh based product will be designed for scaling up to the client-server version. The prototype client server based version will be published by September 1995. When feedback from use of the client-server prototype is provided by local students and others at remote sites, a final network version will be made available by September 1996. Report of the project's efforts and outcomes will be prepared by September 1996.
5. Bibliography of Pertinent Literature :
T. J. Berners-Lee, R. Cailliau, J-F. Groff, B. Pollermann (1992) "World Wide Web: The Information Universe." In Electronic Networks: Research, Applications, Policy. CERN, Geneva, Switzerland. Downloaded off the internet by anonymous FTP from info.cern.ch in pub/www/doc.
T.J. Berners-Lee and D. Connolly (1993) Hyper Text Markup Language. Version 1.0 of the specification generated from the hypertext on the World Wide Web.
J. Conklin, (1987) "Hypertext: an Introduction and Survey." Computer, September, 1987. Published by IEEE.
B. Hann, P. Kahn, V. Riley, J. Coombs, N. Meyrowitz (1992) "IRIS Hypermedia Services." In Communications of the ACM, January 1992, Vol. 35, No.1. pp. 36-51.
F. Kappe, H. Maurer, N. Sherbakov. (1993) "Hyper-G: A Universal Hypermedia System." In Journal of Educational Multimedia and Hypermedia Vol. 2(1), pp. 39-66.
G. Landow (1989) "Hypertext in Literary Education, Criticism, and Scholarship." In Computers and the Humanities, Vol. 23, pp. 173-198. Kluwer, Netherlands.
G. Landow, P. Delany (1990) "The State of the Art." Hypertext, Hypermedia and Literary Studies (Landow and Delany, Eds.).
K. Lewin (1935) "The Conflict between Aristotelian and Galilean Modes of Thought in Contemporary Psychology." In A Dynamic Theory of Personality: Selected Papers of Kurt Lewin. New York and London, McGraw-Hill.
R. Lawler (1985) Computer Experience and Cognitive Development. Horwood Press. Chichester, UK.
R. Lawler, B. DuBoulay, M. Hughes, H. MacLeod (1986) Cognition and Computers. Horwood Press. Chichester, UK.
R. Lawler (1990) "CASE: a Case Analysis Support Environment." In Hypertext, the State of the Art, McAleese and Green (Eds.). Ablex Publishing, Norwood, NJ.
R. W. Lawler (1996) "On the Merits of the Particular Case." In R. W. Lawler and K. Carley, Case Studies and Computing, Ablex Publishing, Norwood, NJ. (1996).
Yankelovich, Nicole From Electronic Books to Electronic Libraries.
6. Vita of Principal Investigator
7. Selected Publications of the PI
7.1 Five Books:
Lawler, R. W. with K. Carley. (1996) Case Studies and Computing, Ablex, Norwood NJ, 1996.
Lawler, R. W. with M. Yazdani, editors. (1991). Artificial Intelligence and Education, Vol. 2. Ablex, Norwood, NJ.
Lawler, R. W. with M. Yazdani, editors. (1987). Artificial Intelligence and Education, Vol. 1. Ablex, Norwood, NJ.
Lawler, R. W. with B. DuBoulay, and M. Hughes and H. MacLeod. (1986) Cognition and Computers. J. Wiley, New York.
Lawler, R. W. (1985) Computer Experience and Cognitive Development. J. Wiley, New York. (Horwood Press, Chichester UK)
7.2 Five Papers
Lawler, R. W. (1989a). CASE: A Case Analysis Support Environment. Proceedings of Hypertext II, York University, England.
Lawler, R. W. (1989b). Constructing Knowledge through Interactions. In Journal of Mathematical Behavior. October 1990, Vol. 9, no. 2
Lawler, R. W. (1989c). Making Jokes and Learning. Humor: The International Journal of Humor Research. Vol.2, #3, 1989. Mouton de Gruyter, Berlin.
Lawler, R. W. (1986) Coadaptation and the Development of Cognitive Structures. Proceedings of the 7th European Conference on Artificial Intelligence, Vol.1. 586-597.
Lawler, R.W. (1981). The Progressive Construction of Mind. Cognitive Science 5. 1-30
12. Statement of the Impact of the Proposed Research
The project addresses a question critical to the success of national efforts to aid the public in better using network accessible information. The essence of the proposal is that one can make sense of interrelations of files in a two step process. In the first step, a person reads the materials of related on-line files (a file assembly). Where appropriate s/he notes relationships between their elements in a journal. Software tools mark the links between elements in the files. When there is need for reorganizing the information, the user takes the second step. S/he proceeds to construct a classical hypertext (e.g. Xerox' Notecards) by extracting and structuring materials from the journal files, using the links created within the file assembly in the first step
Appendix A - Existing Prototypes for this Project
Produced in 1990: Using Hypercard 1.2.5
The Hypercard Society of Mind (Minsky)
With Minsky's permission, the PI prepared a Hypercard version of the Society of Mind. The idea was that the structure of the Society of Mind (300 short, interrelated essays) would benefit from hypertext linking capabilities. More powerful software environments were considered, but the attempt was made to complete a Hypercard version because it was widely available and free to users.
Features: A simple technique was developed to integrate graphics loading as well as text files. The development of indexing facilities to support different ways of approaching the text was completed. These were made flexible so that users could add their own index-terms to those originally defined by the author.
Limitations: The hypertext support within version one of Hypercard was inflexible and difficult to work with. Further work on this project was deferred until public availability of Macintosh system seven and Hypercard 2. Very limited distribution. Some of the materials are now being absorbed in to a CD-ROM version of the Society of Mind being prepared by Voyager for fall 1993 publication.
Produced in 1991: Using Hypercard 2.1
Lives in Progress (R. W. White)
Robert White, one of the leading case study analysts of the post-war period, published a text with this name. When he retired from the Harvard faculty, White gave his case study corpus materials to the Henry Murray Research Center at Radcliffe College. While the PI was visiting there as a Research Scholar, he created a hypertext of White's text and of some of the research corpus underlying his interpretations with prototypes of the multi-file access and linking techniques described in this document.
Features: The essential purpose of this project was to permit analysts to compare public interpretations of case studies with the materials of the corpora themselves.
Limitations: This work is not publicly available because of privacy constraints placed on the material by the owners.
Appendix B - Materials of the Case Study Corpus
The primary objective of the case study (I call it "The Intimate Study") was to produce a corpus which could be analyzed in such a way as to advance our understanding of learning. At the beginning of The Intimate Study, we were four in my family, two parents and two children, Robby, age 8, and Miriam, age 6. The specific objective we followed was to trace in fine detail Miriam's learning for the six months following her sixth birthday (April 9th through October 8th, 1977). This is the core of The Intimate Study; it is extended beyond this core period by later observations. I recorded her behavior in well structured situations and followed her beyond the confines of the computer laboratory with naturalistic observation in the various settings of her everyday world. Miriam was under continual observation for six months. With very few exceptions, the only times Miriam spent away from home were in my company; most of those times were the hours at project Logo's Children's Learning Lab. Her young age, the limits of her world, and my having the time permitted us to share a common world for the period of The Intimate Study. The data collected during this period are grouped into four matters: profiles; sessions; vignettes; and the log.
The Profiles
The profiles are a series of initial and terminal cognitive examinations. Through these, Miriam's capabilities and styles of thought may be compared with data in the psychological literature and with normal education skills through her performance of these specific tasks [1]:
- School tasks: reading and arithmetic skills.
- Piagetian tasks: one-to-one correspondence; class inclusion; continuous quantity; time; object volume; combinations; substance; weight; displacement volume; backspinning a ball; beam balance; multiple seriation.
- Stanford-Binet Intelligence Scale: test L-M at 6 years, 19 days.
Other materials of an idiographic character were also collected
The Sessions
The sessions were mechanically recorded, all on audio tape and many on videotape. These relatively formal, directed working sessions at MIT's Logo lab (approximately 70) and at home (approximately 20) exhibit more than four months of interactions between Miriam and me in a computer-centered environment. The data are as detailed as any critic could wish.
The Vignettes
Richly interpretive, highly subjective, open to error and overstatement, these materials are essentially ephemeral literary constructs whose purposes are to document events in the social world of our family and to connect themes emergent in the more structured data. The vignettes are like snapshots of thinking or short stories that surfaced in the small society of our family. They are based on selective naturalistic observation of Miriam's behavior beyond the range of mechanical recording and in situations where the recording itself would have been obtrusive. I attempted to capture all unrecordable and significant expressions of development Miriam exhibited during The Intimate Study. To the extent that they record observations by an ever-present scientist in the midst of the action, they attempt to elevate anecdotal reportage to the status of naturalistic observation through the claim that they record ALL thematically interesting behavior in those settings beyond the range of mechanical recording. Though imperfect, these data may still be accepted as additional, well-placed pieces in a complex puzzle, patterns in the development of Miriam's mind. Each vignette includes a short sketch of the point of view from which I judged the content significant.
Interpretations
The primary interpretations of this project are seven detailed analytical chapters. Two of those were published earlier as a part of Cognition and Computers. Five were published in Computer Experience and Cognitive Development. Additional analyses on a smaller scale were also published in those two books. Other interpretations are available from publication in various journals, such as various as "Humour" and "The Journal of Mathematical Behavior." Intermediate materials from the period, both unpublished papers and drafts of papers published later are also available for inclusion with the interpretations of the database.
Models and Their Descriptions
Computer based models representing some of the material analyzed in the interpretations is available. The models were originally written in Lisp but can be rewritten as required. This kind of modeling has been described in papers published in proceedings of the Cognitive Science Society and in the European Conference on Artificial Intelligence. The early style of modeling will be extended. It will also be possible to use the same material for modeling with different techniques.
Publication notes:
Text notes: