Time in Interpersonal Media

February 15, 2002

Abstract

This paper explores the ways in which media can be structured to manipulate time. Many media are, in effect, time machines that allow us not only to travel through time, but to manipulate the reality and perception of time in different ways. A number of time-related characteristics of media are discussed, including persistence, speed, turn taking, performance interval, date and time stamps, indexing, asynchronous access, random access, length measurements, reference pointers, linearity, deserialization, reserialization, synchrony, and supersynchrony.  Several time-related characteristics of media, including asynchrony, near-synchrony, desequencing, resequencing, hyperlinearity, and metalinearity, are illustrated with transcripts of instant   messenger and wiki collaborative interaction.

Introduction

Our usual invocation of the term time machine follows H.G. Wells general model as described in his classic novel "The Time Machine". It describes a means of transport that moves people through time in the same way that conventional vehicles move people through space. This article describes a more pedestrian, but considerably more practical, notion of a time machine, argues that we have been building them for at least tens of thousands of years, describes the ways in which we have used these time machines to manipulate and transport ourselves back through time, and demonstrates how new versions of these time machines are allowing us to manipulate time in new ways.

The time machines that will be described and analyzed here are communication media, particularly interpersonal media. This paper will focus exclusively on characteristics of media (Foulger, 1990, 2002) that reflect ways in which time can be manipulated by media. While these characteristics are a function of media design, this paper will not focus on design issues. It will simply note time-related characteristics of media, a set of media that have (or proto-media that may have) these characteristics. It will then look at the ways in which these characteristics affect media use, effects, and practice. In particular, interpersonal transcripts from instant messaging and Wiki collaborative composition will be analyzed.

Background: Sorting Out Commonplace Notions of Time In Media

The discussion of time and its relationship to media is hardly new to the communication literature. Time is a fundamental measurement construct in conversation analysis (Hopper, 1991) and time and motion studies (e.g. Eisenberg and Goodall, 2001 based on work of Frederick Taylor). The passage of time is implied in developmental models of friendships Rawlins (1992) and relationships (Duck, 1990; Knapp, 2003) Time is treated as a major non-verbal code in the study of chromatics and as a culturally dispensable element of language in the Sapir-Whorf hypothesis. Innis (1951) suggests that enforced time linearity is one of the prices of the transitioning from orality to literacy. McLuhan (1963) extends this critique of literacy while enthusing over the capacity of broadcast, telephone, and other electronic media to collapse time and space (McLuhan, 1963). Ong (1982) suggests that electronic media are breaking up linearity of thought with a new secondary orality. Derrida (1973) discusses the relationship of temporality, history, and punctuality to phenomenal experience (following Heidigger and Husserl) and symbol production (presence, difference, and significance). Ong (1982) describes the memory tricks that allow oral historians to serialize and memorize stories that can persist over long periods.

This range of discussions of time and its relationship to communication is substantial, effectively covering all of the most fundamental constructs associated with communication. There are discussions of codes and language, the messages that we create with language, the people who create messages, the media that carry those messages to consumers of those messages, and, among other effects, the relationships that result from and are maintained by such use. Indeed, if these views of time in communication are treated as instances of invented activity within a medium (following Foulger, 2002), we find that each is incommensurate with most of the others.

Foulger describes media as evolving in five discrete spheres of invention: mediators, characteristics, uses, effects, and practices. Mediators are the construction materials with which a medium is created. Characteristics are statements of a medium's capabilities given a set of mediators that have been put together in a particular way. Uses are the purposes for which people adopt a particular medium. Effects are the results, intended and otherwise, of using a medium (most often for a particular purpose). Practices are the behaviors and rules we adopt in order to optimize our success in using a medium for a particular purpose and to minimize the annoyance associated with such use. While these five spheres of invention are interrelated such that changes in one results in changes in the others, they are incommensurate. One can reasonably discuss how mediators enable characteristics enable uses enable effects enable practices enable uses and practices (Foulger, 2002), but one cannot directly compare a practice (veteran's marching in a Veteran's Day parade) with an effect (national pride) with a use (e.g. history) with a characteristic (e.g. performance persistence) with a mediator (e.g. human memory).

Derrida's "history" reflects a ''use'' of a medium. McLuhan's collapse of time and space is an ''effect'' of a medium. The measurement of time in conversation analysis looks at normal practice, which emerges naturally within a medium. Time and motion studies use scientific method to discover optimal practice. Development models of friendships, relationships, and communication competence look for normal patterns of evolving practice, and the Sapir-Whorf hypothesis looks at language as a cultural practice that has, and responds to, cultural effects. Ong's memory tricks (rhyme, rhythm, song, and story templates) are a set of learned practices whose exercise enables more reliable storage in oral cultures. The systematic application of those practices converts them to a mediator, thus enabling new media in which productions have long-term persistence. Among the discussions of time and communication outlined here, only Innis' enforced time linearity, and Derrida's temporality and punctuality can be regarded as statements of the characteristics of media.

As we look at the history of media invention, is apparent that a number of early media, including tribal historians and storytellers, carvings (venus figurines), cave paintings, notched bones and sticks, tokens, and writing on clay tablets were developed with the specific purpose of preserving a record of the past such that important events could be recalled, reviewed, and relived. It is in this "use" of media that we create our first true Wellsian time machines. An effectively remembered and retold story of the past is, in a very real sense, a true time machine that allows us the return to the past through the eyes of those who lived it. Books have, in effect, infinite performance persistence. Hence they make effective time machines. An irony of Wells' "The Time Machine" is that the book is that which it claims to be about, a time machine that carries the reader not only into a recalled past and an imagined future, but into the assumptions and conventions of Wells' 1890's England.

Time-Related Characteristics of Media

Persistence, the characteristic of media that makes, history, among other uses, possible, is only one of a fairly large number of media characteristics that reflect the ability of media to manipulated time. Our traditional notions of interpersonal communication assume a small range of the values associated with some of these characteristics. We would expect that both the minimum speed with which a message can move from a message creator to its audience and the minimum speed with which feedback from the audience can reach the original message creator would be immediate in interpersonal media. Such is certainly the case in face-to-face communication, which is live, real time dynamic, and entails turn taking (all time-time related variables). The usual synchronous group size associated interpersonal media is two (a dyad) or a few (a small group). We would not associate a regular performance interval, prerecorded elements, planning, drafts, date stamps, indexing, asynchronous access, random access, time stamps, message timing or performance length measurements (again, all time-related characteristics) with interpersonal media. In particular, we don't generally associate persistence with interpersonal media. People rarely record their conversations.

These associations are not entirely fair. Notes, personal letters, interoffice memos, and other correspondence media have long histories. Postal systems (the transmission infrastructure of personal letters) has a history that traces back at least 2500 years to The oldest continuous operating postal system (that of China) can be traced back at least 2300 years (Wikipedia, 2003). Notes, personal letters, and interoffice memos are all interpersonal, at least insofar as one person sends a note, letter, or memo to another. Minimum transmission and feedback speeds are relatively slow (minutes for notes, hours or days for memos, and days or weeks for letters). Performances are not live. But they do normally entail turn taking and generally are exchanged within dyads or small groups. But unlike face-to-face interaction and other paradigmatic cases of interpersonal media, they can and sometimes are sent at regular intervals; can and sometimes do contain prerecorded elements (attachments, stickers, rose petals, etc.); can and sometimes do follow a plan (outline). Sometimes they even involve drafts. Virtually all letters include date stamps. Letters can be saved indefinitely, reproduced at will, and even shared with others long after they were written. Saved letters, like books, are time machines that allow us to re-experience the past.

Alternatives to Linearity in Media

One notes that many of these characteristics give specific shape to Derrida's temporality; that characteristics like live and real time dynamic given specific shape to Derrida's notion of punctuality, and that persistence enables Derrida's notion of history. It is time to turn our attention to another time-related media characteristic that is commonly discussed in the literature: linearity. Most discussions of linearity focus on just one of its aspects: sequence. Writing (and all of the media that make use of written language) is often critiqued for imposing sequence on our thinking. Oral societies, we are told, are less likely to fall into the traps of linear thinking because memory allows a more random association of concepts and relationships than a written record does. Ong's description of the memory techniques of oral societies puts the lie to at some of this argument, but it remains that written media are often criticized for enforcing linearity.

This critique may be fair so long as our notion of linearity is restricted to sequence. Books, newspapers, and other documents often have a strongly linear structures that make it difficult to describe alternative logics. The critique can be unfair insofar as all media entail at least some level of a generic sequence that is imposed by the passage of time. No message recipient can receive or reply to a message before it is created, even in face-to-face communication. Creation always precedes receipt. Reply always follows that which is replied to. Performances (including conversations) in many media entail both a greeting ritual that precedes and a closing ritual that ends the performance. Many conversations follow, at least to some extent, rituals and generic scripts that have proven successful in accomplishing specific purposes. It is, perhaps, this experience of logic bound within time that sometimes causes us to confuse logical association and causality.

Time is not the only source of linearity, even if linearity is almost always experienced as a function of time. There are other logics of presentation which are inherent to a wide variety of media. Many of these variations of linearity that have a long history in various media. Recently, networked computers have made additional variations possible: These variations include delinearization, relinearization, desequencing, resequencing, indexing, reference pointers, multilinearity, hyperlinearity, metalinearity, synchrony, asynchrony, near-synchrony, and supersynchrony.  Each will be discussed below.

Delinearization

The non-sequential placement or arrival of messages. While articles in a newspaper will often be organized in sections and there is a notable generic practice of putting national and world news in the first section, local news in the second section, there is no necessary ordering to the sections. Newspaper articles are usually placed within a section according to their importance rather than their relationship to one another. Personal Letters can easily arrive out of sequence, especially when they are mailed from different places or under difficult conditions. Out of sequence arrival is even more common in Internet media. The ordering of articles on Usenet newsgroups hinges entirely on when they arrive at each server. It is rare for different newsgroup servers to retain messages in exactly the same order. ListServes often do better, but here ordering is often decided at the end users e-mail box. There is no guarantee that differences in time zones won't result in messages appearing out of order. On-line databases often entail no necessary ordering of content at all. The user determines selection and ordering.

Relinearization

The placement of messages in a logical order based on time and date sequence. Dramatic and sudden relinearization of events is one of the more common methods through which "deus ex machina" endings resolve themselves in theatrical productions. These relinearizations often repunctuate a series of events for an audience such that actions that were seen as problematic are suddenly seen, based on revelation of new information or a different sequence of events, to be laudable; actions that were seen as laudable suddenly seem problematic. Drastically relinearized content, like that seen in the recent movie "Memento", continuously challenge the audience to make sense of events and responsibilities. Most of us have had the experience of relinearizing a shopping list; of writing down a series of things we need to buy and then checking things off the list out of order as we shop in a store that has been organized on different principles. On-line databases often allow content to be sorted by date, time, and/or other logical sequence (relevance). Agent software enables relinearization of delinearized Internet content (Grosof and Foulger, 1996?).

Desequencing

Even where the linear time sequence of messages is preserved, the logical sequence may not be. Internet media like instant messaging and chat frequently involve interaction sequences in which parallel threads (something we don't generally see in face-to-face or telephone interaction) occur within a single conversation, with references pointing back two, three or more messages in the interaction. Two examples of desequencing will be analyzed below.

Resequencing

With computer media, software agents and/or sort capabilities can automatically reorder almost anything that can arrive out of order. Most e-mail, and computer conferencing interfaces allow messages to be sequenced in a variety of ways. Most allow messages to be sorted by sender, date and time of arrival, subject, and by reference change. Instant messaging agents might easily be written that rethreaded content to keep replies together with the messages they reply to. It is not clear that anybody actually cares about this feature, but the logic required to do it is fairly simple.

Indexing

Indexing of materials allows users to create their own linearity s as the use materials. The concept of indexing is simple and, for the most part logic neutral. The primary logic that is likely to be found in an index is the abstract ordering imposed by alphabetical sequence. Secondary logics can sometimes be observed in gatherings of related sub-entries under a particular index term. There is nothing new about indexing. It is the primary logic by which reference books (dictionaries, encyclopedias, etc.) are created, and they are commonly found at the end of textbooks and other book length scholarship. Indexes allow consumers to ignore the linear logic of a book and turn straight to the material they are looking for. Computers have simplified the process of creating and using index pointers within a medium. Agents can browse electronic texts to automatically create indexes. The primary tool of indexing in computer-based media is search facilities and engines. On-line databases are generally constructed such that fields in the database can be searched across database entries. E-mail, computer conferencing, and web site interfaces and engines often provide similar facilities. Search is a built-in feature of virtually all of the commonly used Wiki collaborative composition engines and sites. Some of the most successful sites on the web are internet search engines that index web pages across the web.

Reference Pointers

Reference pointers allow users to find materials that are referenced within a document. The concept of reference pointers is fundamental to scholarship. We routinely provide pointers to the work we reference in scholarly papers and texts. Footnotes, endnotes, reference lists, bibliographies, and references are all variants of reference pointers that are commonly found in the scholarly literature, with different publication standards favoring different mixes of these elements. Reference pointers do not necessarily imply any logic deeper than "I found material related to this paper discussed here." Footnotes, endnotes, and references generally make stronger statements ("I found this concept discussed here") than do reference lists or bibliographies. Computer-based media enhance the value of reference pointers in a variety of ways. A bookmark list is a personally maintained collection of reference pointers. A hyperlink can point directly to a referenced document and/or a document that explores a concept in greater detail. Relational databases use key fields to tie together otherwise incommensurate tables such that the data in one table can be joined with the data in another. E-mail and computer conferences often provide reference pointers (re: entries within e-mail subject lines and Ref: pointers in computer conference posts such that a reader can follow a discussion thread back from reply to the material replied to. These reference pointers enable some of the kinds of relinearization and resequencing discussed above. They also enable several variants on linearity that are discussed below.

Multilinearity

Multilinearity refers to the use of forward reference pointers to create multiple paths through a set of parallel logics, a related set of ideas, or a document set. Multilinearity is apparent in several book forms that precede the widespread adoption of hypermedia and, in at least one case, the development of computers. The oldest examples are embodied in identification taxonomies in biology, geology and other fields. One can, for instance, identify the a species of bird, while observing it, by serially identifying specific characteristics of the bird, starting with a page one enumeration of a single characteristic. Each identification points to a specific page number in the book where another characteristic is enumerated. A selection on that page leads to another page, and so on until the specific species of bird is identified. In the 1970's a short-lived genre of multilane books, multilane novels, arrived. In these novels the reader would read, in effect a chapter and then make a decision about the stories action. Each decision led to a different chapter, thus allowing the story to develop in multiple directions. These books were, in effect, a print version of computer games like Colossal Cave that were already in use. As computers became increasingly ubiquitous, this multilinear form started to wane. Multilinearity is far more fully developed in story-focused electronic games, multi-user domains (MUDs), and web hypermedia, where several distinct forms can be identified. One, which might be called serial multilinearity, follows the simple model of bird books and multilinear novels by providing a set of choices at the bottom of each page. A second, which might be called abstract multilinearity (abstract has dual and equally valid meanings) embeds the forward links into the text of each page such that a page is an abstraction of a larger set of ideas, each of which is forward referenced by in-line pointers. In this second form one might, for instance, publish the abstract of a dissertation as the top page of a site, with key ideas and materials forward references from the abstract. One notes a real difference in the feel of serial multilinearity and abstract multilinearity. Serial multilinearity provides a parallel set of forward paths through a set of ideas, much as a multilinear novel or a bird book does. Abstract multilinearity proceeds from surface to depth without specifying any necessary linearity. The abstract (top level document) may suggest an ordering with which ideas might be consumed, but the actual sequence with which ideas are consumed is up to the reader. Abstract multilinearity has no real parallel in prior media, although the effect is sometimes accidentally achieved in encyclopedia entries and dictionary definitions. These ideas are often confused in current literature that makes use of the term multilinearity.

Hyperlinearity

Hyperlinearity is a secondary multi-linearity in an individual can create multiple and parallel logics out of the work of others. Hyperlinearity assumes, and builds on, multilinearity. The simplest realization of hyperlinearity is a list, and it is reasonable to note that many of the oldest written documents (e.g. clay tablets) are nothing more than lists that refer to other things. A reading list that guides a student through a set of ideas is an excellent example of hyperlinearity in action. It is usually not difficult to find multiple lists that recommend overlapping sets of readings in different orders, much as one find multiple overlapping lists of the best and worst movies in the critics columns of newspapers at the end of each year. Hypermedia has spawned a variety of these list forms. The bookmark feature of web browsers allows end users to create, edit, and reorder a set of pointers to a set of web pages. While bookmark lists are often unordered collections of a semi-random collection of pages a user is interested in (much as a shopping list often is), it can be, and often is structured into ordered lists of documents within user defined categories. Where these lists are published on the web (this is very close to Ted Nelson's original concept of a home page), others can review those logics and select their own. Amazon.com s "listmania" provides a multiplicity of such lists, with each consumer of lists encouraged to compose their own. Wiki collaborative composition sites enable a self-defining group of individuals to construct such lists collaboratively. More interesting, however, are the hyperlinear documents that collaborative composition web technologies enable.

Hyperlinearity need not be expressed as lists. A more complex and satisfying hyperlinearity can be achieved by presenting an alternative logic that makes abstract or serial multilinear references to the work of others. Web rings in which each site on the ring are an excellent example of serial hyperlinearity. Each site in the ring provides forward references to one or more other sites (one such web ring, the wiki tour bus, provides intersecting alternative "routes" that provide alternative views of better known wiki sites. Abstract hyperlinearity can be achieved in any web page that describes a set of ideas and the web pages of others that detail them. Such abstract hyperlinearity is commonly observed on collaborative composition wiki sites, where cross linking related concepts on other pages is a norm that helps to reduce the noise levels on most pages. It can be argued that hyperlinearity fully resolves, within a written medium, the critique of writing that claims that writing creates and forces some logics in preference to others. Because web-style linking has spread widely into other internet media, however, hyperlinearity is hardly constrained to web sites. E-mail and instant messenger each provide alternate ways to present hyperlinear logics through the referencing of web pages and other Internet content.

Metalinearity

Linearity expresses sequence. The varieties of linearity are under examined in the literature, and the notions of delinearization, relinearization, desequencing, and resequencing are attempts to express some of this variety and the ways in which it can be manipulated. Metalinearity (literally a linearity about a linearity) achieves one or more linearity s by obscuring other linearity s. Such is a commonplace in wiki collaborative composition spaces, in which a series of contributors often erase all, or almost all, traces of interaction sequence while still presenting one or more (e.g. hyperlinear) logics. There is nothing new about metalinearity or the practices that often accompany it. A movie script will often pass through several sets of hands on its path from concept to screen, with each new writer adding, removing, and reordering content. Coauthored books often entail a similar process in which the multiple authors contribute different initial content based on an initial outline and division of labor and the book passes back and forth as the authors smooth and optimize the text. There is generally no way for an viewer or reader to be able to identify the sequence of changes from the end product (movie or book) alone. Wiki collaborative composition turns this metalinearity into an open-ended process in which documents can evolve infinitely based on the work of an open-ended set of self-identifying contributors. There are entries on Wikipedia (still less than two years old) which have been edited by dozens of people. Wikipedia entries still read like cohesive documents, however. The smoothness of Wikipedia pages is not the norm for Wikis. It is more common for metadiscussion of the documents content to be interspersed within the logic of the discussion, as will be the case in the wiki example discussed below.

Synchrony

The notion of synchrony goes somewhat beyond that of linearity, allowing multiple parallel messages to be enacted within the same conversation at the same time. There is nothing new about synchrony. The parallel encoding of non-verbal communication, at least in part as feedback, to the primary and sequentially encoded (via turn taking) verbal message of face-to-face and, to a lesser extent, telephone interaction is synchronous. So are the synchronous two-way messaging capabilities of TCP/IP-based internet protocols. Synchronous internet media include shared whiteboards and shared desktops in which multiple users on multiple machines can work in the same application at the same time. Chat and instant messaging can also be synchronous insofar as the interactants can send messages at the same time.

Asynchrony

Asynchrony is the obverse of synchrony. It presumes that messages cannot be exchanged simultaneously. Asynchrony is commonplace in publishing, art, and most broadcast media, where long production delays (capacitance) ensure that transmission and reception cannot occur at the same time. It is reasonable to characterize asynchrony in two ways, each reflecting a different source of delay in message delivery. The first, delivery time, refers to the amount of time it takes for a specific document to transit from its initial composition to its consumption. The second, capacitance, refers to scheduled delays in which material is collected together such that it can be delivered as a unit. Newspapers and television news minimize both delivery time and capacitance by setting daily deadlines for the arrival and processing of news such that it can be published or broadcast. Both the delivery times and capacitance associated with books and movies, however, can often be measured in years. Sorting, transportation and delivery give postal systems delivery times that are measurable in days or weeks, but with, in general, a daily capacitance. E-mail and computer conferencing delivery times are often measurable in seconds. E-mail will generally entail no capacitance beyond the rate at which an individual sets their mail browser to collect mail from a server. ListServe-style computer conferencing will sometimes enforce a capacitance based on a combination of time (usually daily) and submission rates (which, if high enough, could result in multiple mailings per day).

Near-Synchrony

Internet media have enabled a new variant of synchrony in which the line between synchronous and asynchronous blurs. The normal operation of a computer conference, for instance, is asynchronous. Content is sent to, and read on, a central server or network of central servers. Where the delivery time and capacitance of an asynchronous media approach zero (e.g. entries can be viewed as soon as they are posted), however, it becomes possible for two or more participants who are using the system at the same time to engage in a dialog that feels synchronous. The notion of near-synchrony then is one of creating or experiencing the impression of synchrony in an asynchronous medium. Near-synchrony can be experienced in e-mail when both parties in an e-mail exchange are watching their mailboxes at the same time. Instant messaging is designed around the creation of a near-synchronous experience. An example of instant messenger interaction which illustrates this near synchrony is discussed below.

Supersynchony

In microprocessor design, supersynchronous refers to the ability of a component to operate either synchronously or asynchronously (Bowie, 2000). The same concept can be readily applied to human communication, where it refers to otherwise synchronous communication systems which can be experienced asynchronously when needed. In a supersynchronous conference calling medium a participant could, if interrupted by a person or event not associated with the conference call, ask the computer to "record" interaction while the interrupt was processed and then catch back up by either listening to the recorded interaction at a higher speed, reading an automatically generated voice to text transcript (we can read much faster than we can listen), and by skipping over material that seems uninteresting. This ability to shift, at will, between synchronous participation in a medium and asynchronous lurking, extends an already emerging new interaction paradigm: the synchronous media use ecology. Synchronous media use ecologies were all but unheard of twenty years ago. They have already become commonplace for our students, many of whom interact with their friends face-to-face while simultaneous interacting with remote friends on cellular telephones. I now commonly find students who will have six or more instant messenger conversations going with friends at the same time, and who readily acknowledge taking and making phone calls and participating in propinquitous (call it back-to-back) interaction with their friends at the same time. Supersynchronous interaction extends this paradigm by allowing synchronous interaction to be treated asynchronously and potentially allowing an individual to participate in two or more concurrent synchronous interactions. The author is not currently aware of any supersynchronous media in practical use, but the concept is achievable with technology that is already in hand.

Manipulating Time with Instant Messaging

Each of these alternatives to linearity offers new ways of bending time in media. Some change our experience of time. Others allow us to manipulate, warp, or even obliterate time linearity. This bears illustration, and we will now do a limited conversation analysis of two transcripts: one of an instant messenger interchange and the other of the interaction associated with a wiki collaborative composition page.

Table 1 shows an instant messenger interaction. The interaction, from greeting to the end of the reproduced transcript, is relatively short, with 24 speech acts. It is the beginning of a longer conversation that continues on for another 50 speech acts. The present analysis will presume some familiarity with the well-documented discourse trajectories that have been documented in telephone interaction (Hopper, 1991). Readers who want to check on some of the comparisons made here and in the discussion of wiki interaction below will need to refer to the conversation analysis literature associated with telephone interaction (Hopper, 1991) or face-to-face interaction.

The transcript does not cover some of the richness associated with interaction on AOL Instant Messenger, where this transcript was recorded, or other instant messaging systems. In particular, it does not illustrate the telepresence, via the buddy list, that let user1 know that user2 was available to potentially have a conversation, the instant messenger ID or handle that both identifies a user uniquely and makes a statement of personal identity, or the buddy icons that made a second statement of such identity. It can be noted, however, that both participants had handles that reflected their primary interests and that each had a personally selected animated GIF for a buddy icon. The transcript does not reflect some of the finer nuances of the interaction. The version of instant messenger used by one of the users provided activity information such that a receiver of messages could see that a message was being drafted. (e.g. "user2 is typing"). It should also be noted that both participants were messaging with other users at the same time they were engaged in this interaction. One of the participants was also on the telephone during the discussion. The surrounding synchronous use media ecology is not directly reflected in the transcript, but it does effect the time intervals associated with some of the interacts.

Table 1: An AOL Instant Messenger dialogue. From personal archive. Used with permission.
Speecch Act Topic Replies to Extends Reply Interval Dialogue
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
1
1
1
1
1
1
1
2
1
2
2
1
1
3
1
3
3
3
3
3
4
3 & 4
3 & 4
3 & 4



2
3
2
4
5

5
8
10
9
12

13
14
16
14
18
19

20 & 21
22
23

.
1
.
.
4

.
.
7
.
. .
.
.
.
.
.
.
16
.
.
.
.
.
.



136 seconds
9 seconds
22 seconds
18 seconds
21 seconds

109 seconds
66 seconds
7 seconds
8 seconds
9 seconds
.
24 seconds
66 seconds
8 seconds
15 seconds
62 seconds
41 seconds

27 and 37
21 seconds
18 seconds

user1 (7:25:09 PM): Back from the studio, I see.
user1 (7:26:37 PM): How was it?
user2 (7:27:23 PM): not too shabby
user1 (7:27:32 PM): So you had a good time?
user2 (7:27:45 PM): we still have some mixing to do, but everything's recorded now
user2 (7:27:50 PM): yep
user1 (7:28:06 PM): That's great.
user1 (7:28:49 PM): Mom said you were coming down with something.
user1 (7:29:55 PM): Was the recording engineer happy?
user2 (7:29:55 PM): yeah, I have a cold, but it didn't really affect the recording too much
user1 (7:30:02 PM): That's good.
user2 (7:30:03 PM): yeah, we kept him laughing
user1 (7:30:12 PM): Yes, but was he happy with the music; the recordings.
user2 (7:30:20 PM): we have an amazing acoustic version of falling down as a hidden track
user2 (7:30:26 PM): yeah, I think so
user1 (7:31:27 PM): I thought you might have the double bass up for just that song.
user2 (7:31:35 PM): yep
user1 (7:31:42 PM): A hidden track?
user2 (7:32:44 PM): yeah, its unlisted on the album, but is the last track
user1 (7:33:25 PM): Ah, very good. Is it the only version of falling down, or a second.
user1 (7:33:35 PM): Can you kick SETI?
user2 (7:34:02 PM): its a second version, and sure
user1 (7:34:23 PM): Good stuff, and good stuff.
user2 (7:34:41 PM): indeed

Several manipulations of time linearity are apparent in the transcript. First, the interaction is clearly near-synchronous (despite the fact that activity information was available for viewing). The interval between speech acts is fairly long by the standards of face to face interaction or telephone interaction, where the interval between speech acts is typically measured in small fractions of a second, but fairly short by the standards of e-mail interaction, where exchanges typically take minutes, hours, or days. The shortest reply interval is 7 seconds. The longest, at the beginning of the conversation, is 136 seconds. Despite these speeds, there are occasional flashes of synchrony, as can be seen in speech acts 9 and 10, which arrive at the same time. This kind of synchrony amidst asynchrony is a not infrequent occurrence in near-synchronous media like instant messaging. Note that the level of synchrony observed in this case exceeds that of face-to-face interaction or telephone conversation. A similarly synchronous pair of messages in those media would be labeled a floor fight that would generally be resolved by one speaker or the other taking the floor. Here the simultaneity is barely noticeable and the conversation continues to proceed smoothly.

A second manipulation of time-linearity is apparent as a desequencing of threads within the interaction. Separate threads are pursued in parallel at several points in the interaction, and the last three speech acts actually combine those threads in speech acts that refer to two unrelated themes. This represents a huge departure from the usual practice in face-to-face interaction and telephone conversation, where topics are almost always managed one at a time and replies almost always refer directly to the previous speech act. This practice is so commonplace that conversation analysis generally eschews the study of individual speech acts in favor of interacts, with every utterance considered in combination with the speech act that immediately proceeds it. Here, however, interacts are frequently separated by intervening speech acts. Only 10 of the 20 speech acts that reply to a prior speech act refer to the immediately preceding act in the transcript. The others refer to acts that occurred up to 4 speech acts back. These kinds of back references certainly occur in normal speech, but the 50% rate of multiple lag replies probably requires us to structure interacts around the logic of the interaction rather than time sequence in which the speech acts occur.

There are ways to design instant messaging software such that interaction could be resequenced to preserve logical chain of interacts.  It is not clear, however, that anyone is having a problem following the time sequenced transcripts as they are, and it is unlikely that anyone will fix a problem that nobody cares about.

Manipulating Time in Wiki Collaborative Composition

Table 2 shows a transcript of a Wiki Collaborative Composition page. The page is what is generally referred to as a user page on which an individual describes himself or herself. In this case the user page makes reference to a series of other Wiki pages, and the interaction discusses what at least some members of the Wiki community regard as a violation of the wiki sites unstated rules (at the time of this event, the wiki made the general claim that it had no rules). The web page shown in Figure 1 is not document level smooth, as is the case for many wiki pages. The formatting of the page, the sequence ofinteracts, and their relationship to each other clearly illustrate that interaction has occurred. Except for the obvious replies, however, it is not at all clear what the time relationship of the 12 speech acts described in Table 2 is. Indeed, it should be noted that there are actually more than 12 speech acts. Several speech acts where deleted in whole or part in the course of this interaction. It should also be noted that the interaction here is entirely asynchronous. Indeed, anything approaching synchrony in this interaction might have accidentally erased immediately prior speech acts.

Table 2: An interactively constructed Wiki page. Note that the able format used here breaks up what is actually a continuous document into discrete speech acts. These speech acts are only differentiatable, on the actual wiki page, by the selection of font, which is preserved here. Viewable, in its latest version, at http://www.c2.com/cgi/wiki?DavisFoulger. Used with permission of the page owner.

Davis Foulger

1
When I first arrived at this Wiki, I wrote a nice description of myself. That description has since migrated to my own Wiki. I still play here, however, so if you have an interest in my relationship to the computer industry, you can find far more detail of it than anyone might ever want to read in my hypermedia resume: http://evolutionarymedia.com/davis/resume). More interesting info, including a pleasant Java game, can be found at my personal home page (http://www.davis.foulger.net). My Oswego State University home page (http://www.oswego.edu/~dfoulger) has info on the classes I'm currently teaching. My SetiAtHome (see also DistributedComputing) Stats are at http://setiathome.ssl.berkeley.edu/cgi-bin/cgi?email=davis_foulger%40prodigy.net&cmd=user_stats_new. My business home page (http://www.evolutionarymedia.com) has some inadequate information about my fledgling consulting business. 10

I have now (as of October 9) set up my own Wiki and moved the pages that have caused such controversy here (MoralQuestionInMedia and SuicideOnChildrensTelevision?). I don't expect my Wiki will be terribly interesting to folks here, but if you want to come and play, it is currently located at http://evolutionarymedia.com/cgi-bin/wiki.cgi (note that the location could easily change). I have, with Ward's blessing, modified WelcomeVisitors somewhat to make experiences like mine less likely. I have also suggested, in DiscussionOfCategories, a new category, CategorySandbox?, that might be used to reduce the impact of errors like this ones I've made here.-- Davis Foulger
11

10

A group of my friends and colleagues may be using this page: DavisFoulger to discuss the ethics of polluting a focused, long standing wiki community with a bunch of irrelevant pedagogical rants when you already have your own server where you could host them and you could easily run up one of the myriad free WikiClones there instead ... -- PeterMerel.
3

I'm sorry you and your friends feel that way. I took the WelcomeVisitors page at its word, and have created three pages, including this one, as experiments while I worked on setting up my own Wiki. I haven't polluted other pages on this Wiki, and I don't think I created any of the rants associated with the ones I did create (except maybe now on this page). The second page I created, MoralQuestionInMedia, was intended to demonstrated the power of Wiki as a collaborative medium to a group of scholars. I had proposed that we (the group of scholars) set up our own Wiki for the purpose of collaboratively developing Ethical Cases across a variety of fields. The demo, which started out as about 100 words in WikiWikiSandbox, did a pretty good job of demonstrating that value, and the "rants" made a useful contribution to the demo. I moved it to a new page from Sandbox so the demo would be preserved for continuing use over the next week or so. The second page, SuicideOnChildrensTelevision?, was intended to solve a specific problem some students were having this week. My apologies.
4
I have since moved them to my own Wiki. -- Davis Foulger 12

[Note that Peter has his own wiki! Try using 'minor edits'. Also see WriteNewPages.]
5

Thanks for prodding me into exploring further. I really didn't want to like Peter after his threatened gang violence against my trivial little experiments in Wiki use. I was ready to push him off in a corner as just another software guy who has made up a weird and pretentious title for himself, but now I've read VeryGoodSeats, and I like Peter. I'm sure that, if we met on an airplane, we would share witty puns and discuss some interesting backwaters of language choice. Heck, I've made up weird and pretentious titles for myself too.
7
You like me?! Now the gang is going to have to really rough you up! Ward, hand me the tire iron. Um, Ward? Hey, fellows? Gang? Hey, wait up, where are you all going? --PeterMerel hulking off striking surly poses and kicking cans. 8

:-) -- Davis
9


Davis, after a week away from Wiki I realized when checking RecentChanges tonight that I must have missed all the excitement! I have also found on Wiki that discussion of certain topics - especially programming, music, and science fiction - is pleasant and agreeable, but discussion of controversial topics leads to more heat than light. I'm certainly a newcomer to Wiki myself and can't speak on anyone else's behalf, but I am dismayed to learn of the negative reception your new pages received. Please let me know if you will be accepting non-academic visitors to your Owsego wiki when it is set up. -- ChrisBaugh, Sat 10/6
5

Thanks, Chris. A nice comment about the value of the "minor edits" button has calmed me considerably. Controversial topics do create interesting problems for discussion groups, and there are elements of Wiki "style" that may actually make Wiki a more difficult place to have controversial discussions. It certainly appears to push discussions towards a format that I've referred to as polar debate in previous research. I'll let you know on the scope of the Oswego site after I have it set up, but its nice to know that someone may be interested. Practically speaking, there may not be much I can do to keep folks out, and I can't think why I would want to, but first I have to set it up, and given my recent experiences with Windows and Sun-based CGI serving, the first step will probably be to set up a Linux box.

6

CategoryHomePage
2

Note that the table format used in Table 2 breaks up what is actually a continuous document into discrete speech acts. These speech acts are only differentiated, on the actual wiki page, by the selection of font, which is preserved here. Viewable, in its latest version, at http://www.c2.com/cgi/wiki?DavisFoulger. Used with permission of the page owner. The first thing to notice in this transcript is the obvious desequencing of interacts. The remains of the first speech (e.g. the title of the page) and the entirety of the second speech act (which actually followed several now invisible additions to the first interact) are at the top and bottom of the page. The third to eleventh speech acts appear in between. The eleventh speech act is actually the third speech act visible on the page. It splits the tenth speech act, which appears in the second and four positions in the table. The remaining speech acts appear in the following order: 3, 4, 12, 5, 7, 8, 9, 5, and 6. Notice in this sequence that while there are several serial chains (e.g. true interacts), including 3 to 4, 7 to 9, and 5 to 6, that there are several out of sequence elements. Speech act 12 appears between speech acts 3 and 4. Speech act 5 is actually split. It is initially replied to by speech act six. Later, a portion of it is replied to by speech acts seven to nine. Time sequence, in this interaction, is broken and can never be fully recovered. Resequencing is not a possibility. The logic of the page overrules the linearity of time.

The second thing to notice is the various uses of reference pointers on the page. Speech act 10 (which partially reconstructs much of interact one) points off to a variety of sites that are unrelated to this site. All relate usefully to the user who has created the user page, but they express no particular sequence beyond "here are some things to know about me. This is, for all intents and purposes, a list embedded in prose. Speech act 11 offers a similar reference pointer.

Some of these uses of reference pointers, most notably those of speech acts 3, 4, and 11, are clearly attempts to construct hyperlinear logics. Speech act 3 is, in effect, a hijacking of the user page in which a long standing member of the wiki criticizes the user for misusing the wiki, with supporting links supplied. What doesn't show is that speech act 3 actually replaced the prior contents of the page, which were subsequently invisible. Speech act 4, by contrast, attempts to justify the perceived misuse by explaining the purpose of a series of pages. Speech act 11 attempts to construct another multilinearity by showing how some good may have come from the discussion, again with reference to supporting pages.

Other speech acts, most notably acts 5 and 7 clearly express metalinearity. Both entail metadiscussion of the immediate page, and include reference pointers that support the perspective expressed.

The transcript of Table 2 is, by any standard, a wild ride, but most interacts are signed (typical for wiki interaction). Despite the fact that the ordering of the discussion is opaque, the result is a real transcript that reads something like an interpersonal interaction. There just isn't any sense of the true time linearity within which it was created. Some wiki pages retain a smoothness despite this kind of interactive modification of the text. Others are a lot like this example, with interleaved interactive fragments not unlike those associated with a quote/counter quote style of e-mail or computer conference reply (Foulger, 1990), but with a sequence can only be inferred from the context and may be ever be teased out entirely.

Conclusion

This paper started with the assertion that communication media can be usefully treated as real world time machines which, through their varied characteristics, enable a variety of useful manipulations of time and our experience of time.  The notion of time machine is fairly pedestrian by comparison with fictional representations of time machines that physically transport people through time, but is substantially more practical.  Practical communication-oriented time machines have been around for a long time, but new computer-based media allow us to manipulate time in new ways.  Indeed, as we can see in instant messenger and wiki collaborative composition, it is clear that time linearity can be significantly warped in the former and obliterated in the  latter. 

Interaction without time-imposed interaction structures is something different in our experience of interpersonal communication and these new media have attracted large followings of users. Instant Messaging has become a medium of choice, especially among students, and continues to attract a wider audience. Wiki collaborative composition sites are also growing rapidly, and it can be expected that additional Internet media with even more extreme characteristics (supersynchrony, for instance) will emerge in the near future.  It appears that there may be considerable value in taking a closer look at the ways in which media can be constructed to manipulate time.

References