Derek J. Smith
Site first loaded 11th January 2000. This version 09:00 BST 21st May 2008 [misc. updates]
![[mugshot]](http://www.smithsrisca.demon.co.uk/mugshot.JPG)
Senior Lecturer
in Informatics and Cognitive Science at University
of Wales Institute, Cardiff
CONTACT DETAILS:
AT A GLANCE
I graduated as a psychologist in 1972 and spent most of the next two decades in commerce and industry. This included lengthy periods firstly as a small businessman and then as a senior analyst-programmer with the data processing arm of British Telecom, where I specialised in the design and operation of very large databases and management support systems. Since 1991, I have lectured in both IT and cognitive science at University of Wales Institute, Cardiff (UWIC). During that time, I have been responsible for the cognitive psychology, applied cognitive psychology, psycholinguistics, and neuropsychology modules of the BSc (Hons) Speech and Language Therapy and BSc (Hons) Psychology undergraduate programmes, as well as for the Informatics and Project Management module of the MSc Interprofessional Studies. My research interests include the nature of biological memory mechanisms, the evolution of the mind, and the use of computer animation and Internet technology as teaching methods. My consultancy specialisms include data modelling, systems testing and IT project assurance, cognitive modelling in general, and the cognitive science of the learning experience in both academic and organisational settings in particular.
CURRENT
INTERESTS AND PROJECTS
Coming Soon
The 22nd International Conference on Computational
Linguistics [Coling 2008]
On Sunday
17th August 2008, my colleague Kathryn M. Livesey and I will be presenting a
three-hour tutorial entitled "Data Analysis for Lifelike Computational
Linguistics". This
tutorial will assist cognitive scientists in general, and especially those
interested in human language processing, to get to grips with the distribution
of biological memory in the mind (and thereby to computerise
same more effectively). The tutorial will briefly review the main genres of
cognitive diagramming, and show in practical easy-to-grasp steps how to prepare
a data model of a complex information processing system. Each component skill
will be instructor-supported, and illustrated by extracts from the data model
supporting the Konrad
artificial consciousness system.
More about Konrad [Press Release, 12th March 2008]
Widening Access Provision Series
The following PowerPoint presentations were
inspired by the Welsh Assembly Government's Reaching
Higher-Reaching Wider
initiative for the modernisation of Welsh post-16
education. The first two were screened under the FIRST Campus logo at the 2005 Christmas Lecture of the
South Wales Branch of the British Computer Society, Cardiff Bay Techniquest, 1st December 2005.
|
|
|
|
The Turing Test of Machine Consciousness |
|
|
Promoting Selfhood in the Learning Disabled Child |
|
|
All About Your Brain |
click here NEW!! |
The 2005 Techniquest event also introduced a card sorting version of
the famous Turing Test. Simply prepare a pack of laminated cards, each with a
couple of lines of dialogue on it. Half the cards should use dialogue taken either
from the Turing Test literature or from your personal interaction with one of
the Turing Test software products available online. The other half should use
more or less similar human-human dialogue [the extent of similarity depends on
the age and experience of the proposed audience]. The card pack is then studied
by the contestants and the cards sorted onto one of two receiving trays -
"human-machine conversation" and "human-human conversation"
as they deem appropriate. Discrete markings on the back of the cards allow
rapid scoring right or wrong, and "winners" [you set your own pass
mark] are rewarded with an achievement certificate. Good luck.
National Science and
Engineering Week Series
The
following PowerPoint presentations were inspired by the British Association for
the Advancement of Science's [see their
website] annual National Science and Engineering Weeks [tell me more].
|
The Bilingual Brain (Why Machines Can't Translate for Toffee) [NEWI, March 2008] |
click here NEW!! |
Organisational Cognition and Human Error Series
In 2001, I produced a series of distance learning
resources on the role of human error in disasters. I also delivered a module on
organisational communication at the Asia-Pacific Institute of Business,
|
|
|
|
IT Project
Management Disasters |
|
|
Experiential Learning: The Knowledge Structures and the Cognitive Processes |
|
|
Systems
Thinking: The Knowledge Structures and the Cognitive Processes |
|
|
Mode Error in
System Control |
|
|
Situational
Awareness in Effective Command and Control |
|
|
Military
Disasters (including the battles of New Orleans and Isandhlwana
Hill, the Charge of the Light Brigade, Custer's Last Stand, and the tragic
story of USS Vincennes vs Iran Air Flight
655, 1988) |
|
|
Transportation
Disasters - Aerospace (including the |
|
|
Transportation
Disasters - Maritime (including the Titanic and Exxon Valdez
disasters) |
|
|
Transportation
Disasters - Rail |
Artificial
Intelligence / Machine Consciousness Series
|
|
|
|
Short-Term Memory Subtypes in Computing and Artificial Intelligence: Part 2 - A Brief History of Computing Technology, 1925 to 1942 |
|
|
Short-Term Memory Subtypes in Computing and Artificial Intelligence: Part 3 - A Brief History of Computing Technology, 1943 to 1950 |
|
|
Short-Term Memory Subtypes in Computing and Artificial Intelligence: Part 4 - A Brief History of Computing Technology, 1951 to 1958 |
|
|
Short-Term Memory Subtypes in Computing and Artificial Intelligence: Part 5 - A Brief History of Computing Technology, 1959 to date |
|
|
Short-Term Memory Subtypes in Computing and Artificial Intelligence: Part 6 - Memory Subtypes in Computing |
|
|
Database Navigation and the IDMS Semantic Net |
|
|
Short-Term Memory Subtypes in Computing and Artificial Intelligence: Part 7 - Memory Subtypes in Cognitive Science |
UNDER CONSTRUCTION - first tranche due eventually |
|
Mental
Philosophy Glossary (a multi-file navigable data dictionary on the theme of
self and consciousness - how they are, how they fail, and how they one day
might be programmed onto a machine) |
click here NEW!! |
|
The Konrad Machine
Consciousness Project (some scoping notes) |
click here NEW!! |
Informatics and
Project Management Series
|
|
|
|
Basic
Software and Hardware Concepts |
|
|
Functionality |
|
|
System Defects |
|
|
Strategies
and Platforms |
|
|
System
Feasibility |
|
|
IT Project
Management |
|
|
Going Live |
|
|
Operations,
Maintenance, and Audit |
|
|
Child Protection on the Internet: Quality Systems or Convenient Myth [AS EDITOR, ON BEHALF OF
THE |
|
|
Managing
Small e-Systems: The Secret Skills of Successful Website Design |
Brain and Behaviour Series
|
|
|
|
Microanatomy of
the Cerebral Cortex |
|
|
Communication
and the Naked Ape |
|
|
The Motor
Hierarchy |
|
|
Motor
Programming |
|
|
Biological
Cybernetics |
|
|
The Mind of the Rat [CONFERENCE POWERPOINT] |
Robotics, Cybernetics,
and the Like
|
|
|
|
Basics of
Cybernetics |
|
|
The Eckert-von
Neumann Machine |
|
|
Introduction
to Systems Theory |
|
|
Shannonian Communication
Theory and Biological Communication |
|
|
An
Introduction to Data Modelling for Semantic Network
Designers |
|
|
An
Interdisciplinary Validation of the Ego Defence of
Splitting
[CONFERENCE
POSTER] |
|
|
Freud as
Real-Time Programmer [LECTURE
POWERPOINT] |
click here NEW!! |
Historical Cognitive Models
Series (In Timeline Sequence)
|
|
|
|
Bell-Magendie (1811) - The Anatomist's View |
|
|
Wernicke (1874) - The Aphasiologist's View |
|
|
Kussmaul (1878) - Early "Cog Neuro" View |
|
|
Charcot's (1883) "Bell" - Early "Cog Neuro" View |
|
|
Lichtheim's (1885) "House" |
|
|
Grashey (1885) |
|
|
James (1890) |
|
|
Freud (1891) |
|
|
Wundt (1902) |
Control Hierarchy
Models Series (In Timeline Sequence)
|
|
|
|
Freud (1900) - The Psychoanalyst's View |
|
|
Freud (1923) - The Psychoanalyst's View |
|
|
Freud (1933) - The Psychoanalyst's View |
|
|
Craik (1945) - The Ergonomist's View |
|
|
Wepman et al (1960) - The Clinician's View |
|
|
Frank (1963) - The Information Scientist's View |
|
|
Reader (1969) - The Roboticist's View |
|
|
Dennett
(1978) - The Philosopher's View |
click here NEW!! |
|
Rasmussen (1983) - The Forensic Ergonomist's View |
|
|
Allport (1985) - The Distributed Semantics View |
|
|
Norman (1990) - The State-of-the-Art View |
|
|
Smith (1993)
- The System Analyst's View |
Theoretical Cognitive
Science
|
|
|
|
Speech Errors, Speech Production Models, and Speech Pathology |
|
|
Dyslexia and
the Cognitive Science of |
|
|
Neuropsychology/Aphasiology Timeline |
|
|
Neuropsychology/Aphasiology Glossary |
|
|
Memory Glossary |
|
|
Psycholinguistics Glossary |
|
|
Rational Argument Glossary |
|
|
Supervenience
Glossary |
|
|
History of Human Writing Systems Glossary |
|
|
On database keys, with an application to the Praxisproblem. [CONFERENCE POWERPOINT] |
|
|
The problem of context in sentence production -
Surely a case to re-convene the Data Base Task Group? [CONFERENCE POWERPOINT] |
|
|
How ideas evolve into speech - A computer animation. [CONFERENCE POWERPOINT] |
|
|
Speech Acts - What They Are and How They Can Be
Traumatized by Childhood Sexual Abuse.
[CONFERENCE POWERPOINT] |
|
|
A Gentle Introduction to ..... Frith,
Rees, and Friston's (1998) "Forward
Model" of Self [LECTURE
POWERPOINT] |
Click here
NEW!! |
Modern
Psycholinguistic Models Series (In Timeline Sequence)
|
|
|
|
Gough (1972) |
|
|
Marshall and Newcombe (1973) |
|
|
Morton (1979) |
|
|
Morton (1981) |
|
|
Ellis (1982) |
|
|
Roeltgen and Heilman (1985) |
|
|
Crosson (1985) |
|
|
Ellis and
Young (1988) |
|
|
Kay, Lesser, and Coltheart (1992) |
|
|
Coltheart,
Curtis, Atkins, and Haller (1993) |
Perception, Attention,
and Memory Models (In Timeline Sequence)
|
|
|
|
Broadbent (1958) |
|
|
Sperling (1960) |
|
|
Sperling (1963) |
|
|
Sperling (1967) |
|
|
Pribram's Holonomic Theory of Memory (1969) |
|
|
Atkinson and Shiffrin (1971) |
Occasional Papers and
General Student Resources
|
|
|
|
Codes and Ciphers in History, Part 1 - To 1852 |
|
|
Codes and Ciphers in History, Part 2 - 1853 to 1917 |
|
|
Codes and Ciphers in History, Part 3 - 1918 to 1945 |
|
|
"Peer to Peer" Telecommunication |
|
|
Antidromic Neurotransmission in Modular Processing |
|
|
The Mind of
the Rat [CONFERENCE POWERPOINT] |
click
here NEW!! |
Animated Cognitive
Models Projects
The project before last (Jan-Mar 2002) was to
produce an engineer's analysis of the principles of mental data transmission.
This was presented on 9th April 2002 as a poster-plus-CD-ROM at the
"
PROFESSIONAL
ACTIVITIES
I maintain my professionalism as a computer person by
active involvement with the
RESEARCH
AND TEACHING SPECIALISMS
I am a keen cognitive modeller,
and teach practical modeling skills at a number of points in the psychology
curriculum. My own six-module modular diagram of cognition (Smith,
1993c; Smith
and Stringer, 1997; Smith,
1999c; Smith, 2000b) attempts
(a) to locate, and (b) to show the relationship between, the major types of
long and short term memory in a three-level biological control hierarchy.
CREATIVITY,
PROBLEM SOLVING, AND HUMAN ERROR
During my years as an analyst-programmer, I took part
in many major system testing exercises. This experience has given me a deep
insight into the occasional fragility of human cognition in an increasingly
technical world. As a result, when I was designing my module in applied
cognitive psychology in 1997 I made effective problem solving and human
error two of the four main study units (the others being mathematical
cognition and dyslexia). Students are encouraged firstly to broaden their
theoretical grasp of cognition, and then to apply this knowledge to the real
world. With problem solving, this means looking at the structures and processes
underlying reasoning and decision making, and with human error it means looking
at the cognitive demands of such things as technical design, production quality,
and complex control. These exercises are informed by, and then tested against,
case studies of ineffective problem solving, sloppy system design, and operator
error of all sorts.
The third study unit in my applied cognitive psychology
module is mathematical cognition. This is one of two educationally relevant
study areas (the other being dyslexia), and it brings together (a) theories of cognitive
development, (b) epidemiological data on difficulties and delays in acquiring numeracy skills at school, (c) case study data on the
selective loss of numeracy skills following brain
injury or disease, and (d) mainstream cognitive theory. Because it weaves
together so many disparate threads, mathematical cognition is a good example of
the "cognitive neuropsychological" approach to psychological theorising.
I have lectured in cognitive neuropsychology
since 1991, and cover all aspects of memory theory, including the amnesias and
the underlying biochemistry. In addition, my input into UWIC's
Speech and Language Therapy programme has always
demanded extra focus on the issue of phonological encoding (indeed, it was this
need which eventually prompted my animation of the Working Memory Model of
Memory). My own research into the distribution of phonological ability in the
UK undergraduate population leads me to suspect that there might be a sizeable
minority of students with undiagnosed learning difficulties. From a theoretical standpoint, I favour
the Lashley-Pribram distributed memory approach
(having been impressed with Karl Pribram's holographic hypothesis ever since it first appeared
back in the 'sixties). I also closely follow research into "second
messenger" neurotransmission, because I hold the mechanisms of medium term
neural sensitisation to be fundamental to the
emergence of all complex cognitive faculties, including abstraction,
association, self-awareness, and consciousness. For a general flavour of how I approach memory theory see Smith
(1996b) or Smith (2000b), and
for more on the importance of the biochemistry see CODASYL
Principles in Biological Memory.
CODASYL
PRINCIPLES IN BIOLOGICAL MEMORY
I was particularly fortunate during my years with
British Telecom to work on the design, development, and support of a very large
CODASYL database system. The product in question was ICL's
IDMS(X), a UK-licensed variant of Computer
Associates' CA-IDMS system. This
type of database organises its contents into an
intricate network, rather than into the tabular columns and rows used by more
simplistic systems, and in order to manage the resulting data networks the
system software relies upon a number of clever internal tricks. When I took up
cognitive science in 1991, I decided to look for the biological equivalents of
these mechanisms (after all, the mind had so often been described as a
biological database, that it seemed reasonable to enquire after its database
internals). I eventually concluded that the most compelling similarity was
between the IDMS concept of "database currency" and the biological
mechanisms of second messenger neurotransmission. Both allow their respective
systems to maintain a particular mental theme across a timespan
larger than the span of the immediate here and now, both do this by holding
material momentarily somewhere between short term memory and long term memory,
both combine storage and retrieval functions, and - above all - both exist to
help "bind" widely scattered memory fragments into logical wholes.
What CODASYL databases give us, therefore, is a tangible paradigm for
biological consciousness in general, and a working example of a system
architecture which has successfully overcome the "binding problem" in
particular. For the precise argument and parallel worked examples see Smith
(1997a) and Smith
(1997e), for the paradigm's
utility in addressing the explanatory gap see Smith
(1998c), for the front runners
in mapping the human knowledge network see Doug Lenat's
CYC
Project, for an introduction
to the binding problem consult Valerie Gray Hardcastle's
Association for the Scientific Study of Consciousness website, and for an update on work in progress with a
practical software simulation see Konrad, the Project.
TELECOMMUNICATIONS
PRINCIPLES IN NEURAL TRANSMISSION
I am also interested in how telecommunications
principles might be at work in biological cognition. I began this search by
considering what are known in the trade as network protocols. These are
internationally accepted codes of practice governing such things as (a)
authenticating which stations should actually be in a communication network,
(b) routing messages appropriately, (c) governing when, and at what speeds,
stations are allowed to transmit and receive, (d) arranging for data
compression and decompression, and (e) detecting errors in transmission and
arranging for their correction. There are many highly technical protocols in
force, published by such agencies as the International Standards Organisation, but to get good psychology you actually have
to go no further than the Open
Systems Interconnection (OSI) Model. This is a reasonably non-technical overriding guideline, and what it
recommends is a seven-layered analysis for any given communication, with the
"clever" bits at the top (Layer 7), the physical channel (the wire or
whatever) at the bottom (Layer 1), and a series of vertically organised internal layers in between (Layers 2 to 6).
Now this is
exactly the same arrangement as that found in the sort of hierarchical systems
encountered in biological cognition. True there may be a certain amount of bickering (a) over exactly how
many layers are involved (some authors identify three layers, others five, and
others seven) and (b) over which way round to draw them, but what they all have
in common is that they put the highest form of cognition at the top, keep it
informed via an ascending perceptual hierarchy, and have it act upon the world
via a descending motor hierarchy. This gives us two separate but entirely
complementary vocabularies to describe what happens when mind "A" has
an idea which it wants to communicate to mind "B". In
psycholinguistic terms we would speak of that idea being progressively
converted into sound waves by passing it down through A's motor hierarchy and
then progressively converted back again as it passes up through B's perceptual
hierarchy, whilst in telecoms terms we would speak of a message being
progressively encoded for transmission by the Layer 6 to Layer 2 processes at
Station A, transmitted across the intervening space, and then progressively
decoded by the corresponding Layer 2 to Layer 6 processes at Station B.
Question: So why bother? Why go to the trouble of learning two
vocabularies when you would be forgiven for thinking that one would do?
Indirect
Answer: Because one of the most
refractory aspects of the entire consciousness problem is that the bulk of the
two processing hierarchies - the processes of putting thoughts into words and
words into thoughts - is almost exclusively inaccessible to the
conscious minds involved. Cognition is like an iceberg, in other words - only
the conscious tip of it is visible, supported by a vast invisible storehouse of
everything from motor skills to long term memories. And it is actually quite
disturbing that so much of our short term decision making seems to go on in the
invisible (ie. unconscious) bit, because eventually
one is forced to start questioning the very concept of a human free will (I
strongly recommend Jeffrey Gray and Chris Frith on this subject).
Direct
Answer: Cognitive science
would benefit from the second vocabulary because telecommunications is
expert at the sort of processes which might be involved in preparing
consciousness to become conscious! It has been looking at the hidden parts of the
communication iceberg for 200 years!
For the full
argument see Smith (1997f), for the history of the feedback concept see Smith
(1997g), for an example of
what looks suspiciously like a full duplex telecommunications link in a cognitive model see Coltheart, Curtis, Atkins, and Haller (1993), and for a possible arrangement of three full duplex telecommunications links at the interface between our consciousness
and what lies beneath it see Smith
(1999c) and Smith (2000b).
By describing graduateness
as the extent to which students can be left to their own devices (see
previous), I actually did little justice to the true complexity of this topic. Graduateness is nothing less than a highly complex blend of
a large number of cognitive and behavioural skills.
As to precisely what these skills are, there is no final consensus, but I am
particularly impressed with the list produced by a Quality in Higher Education
study carried out in the early 'nineties. This asked employers to rank 62
separate graduate skills and abilities according to their practical value
within commerce and industry. The top scorers were such factors as willingness
to learn, reliability, self-motivation, teamwork, oral and written
communication, problem-solving ability,
literacy,
and numeracy. Specialist factual knowledge, on the other hand -
the thing most examinations set out to measure - was rated a paltry 59th out of
the 62 attributes! There are therefore three basic issues under debate, (a)
what graduateness is, (b) whether UK graduates have
enough of it, and (c) what to do about it if they do not.
KNOWLEDGE
UPTAKE IN LARGE GROUPS
One of the reasons education often has difficulties
delivering improvements is that it actually has no reliable method of
evaluating its own effectiveness. The most obvious single measure is student
performance, but the assessment process - because it has to be demonstrably
fair - is slow and cumbersome. Even so, there is at least a good common
understanding of what needs to be known and a standard repertoire of methods to
find it out, which is more than can be said for the more introspective issues -
the wisdom of a particular educational initiative, for example, or the relative
effectiveness of competing methods of delivery - where there are few benchmarks
to work to and little consensus as to what should be measured and how it should
then be interpreted. Again I wondered whether there was anything cognitive
science had to offer. Was there, perhaps, a method of evaluating the
effectiveness of the knowledge transfer process (a) at group level, and
(b) without a major data collection exercise being added to the already heavy
burden of educational administration? And again, to cut a long story short,
there was - all you needed to do was track a set of deliberately complicated
messages as they made their way around a representative sample of the group in
question, and see how far they got.
We are, of course, already fully familiar with this
type of message transmission, because it is what we used to do as children when
we played the game of "Chinese Whispers". And yet in psycholinguistic
terms, Chinese Whispers is far more than just a children's game, because
what it is actually doing is probing the hearing, the vocabulary, and the
common understanding of those who pass the message. It follows that if you
factor out the hearing problems by making it a written game with written
messages, the test becomes solely one of the underlying knowledge, and thus
ideal for the rapid evaluation of the teaching process. I trialled
this test paradigm with a group of third year psycholinguistics students in
1996. They were tested on sentences relating to a two-hour neuroanatomy
lecture from the previous year, and their performance compared with a control
group of non-biology lecturers. Working with a five-person chain, student
performance levelled off at about 33% accuracy,
whilst control performance levelled off at about 25%
accuracy, with the 8% advantage deriving from the original lecture input.