Home
| Databases
| WorldLII
| Search
| Feedback
New Zealand Yearbook of New Zealand Jurisprudence |
Last Updated: 24 April 2015
The Role of Computers in
Judicial Reasoning and Analysis
Ian Iredale*
i. intrOdUctiOn
Judgments are traditionally written and delivered in narrative format and
published in volumes of law reports. Occasionally judgments
are delivered live
via television.1 Judgments are also stored in electronic format and
can be accessed electronically online.
Computers, in this age of digital technology, are being used by students,
practitioners and judges in legal research, reasoning and
writing. From word
processing to researching legal databases, accessing law primary and secondary
materials cutting, pasting and
downloading and hyperlinking. However, computers
have much more to offer.
Aim
The aim of this paper is to highlight a number of ways in which computers may
be used to further enhance judicial reasoning and analysis.
Flowcharts
Flowcharts are employed in many disciplines, including engineering, physics
and management. They are coming to find increasing application
in Law.
Flowcharts can be constructed in hard copy using pen and paper. However, there
are computer programs available that are more
effective in constructing and
manipulating flowcharts than doing it manually. Given the increasing complexity
of factual situations
in legal proceedings and provisions in statutes,
flowcharts are being used more and more to represent the facts or to interpret
statutory
provisions.
Flowcharts and diagrams can be found in statutes, case law reports and law
textbooks. For example, diagrams and flowcharts can be
found in statutes such
as, the Income Tax Assessment Act 1997 (Cth) and the Patents
Act
* School of Law, University of Western Sydney.
1 I am not aware of any judgments being delivered directly
online.
1990 (Cth).2 Flowcharts can also be found in textbooks, such as,
Turner’s, Australian Commercial Law and Woellner et al., 2005 Australian
Taxation Law.3 Diagrams also appear in case law reports to assist in
explaining detailed fact situations.4
Flowcharts have been employed to assist students studying taxation law.
Taxation Law is one of the most difficult subjects to teach
and learn, because
of the volume and complexity of the subject matter. Few academics are competent
or willing to teach the subject.
Although income tax law is extensive and complex its foundation lies in two
simple equations:5
Tax payable = Taxable income x Tax rates – Tax offsets and,
Taxable income = Assessable income – Deductions
It is possible to take both of these equations as the basis for a flowchart
or tree diagram with the limbs being the various categories
of income and
deductions. Each limb has a number of branches with “leaves” or
topic boxes that further explain the particular
category of income or deduction.
In hard copy format the flowchart measures 55cm x 85cm and comprises
approximately 230 topic boxes.
A benefit of the flowchart is that it provides an
overview of income and deductions and acts as a roadmap, guiding students
through
the first
2 See for example:
Income Tax Assessment Act 1997 (Cth), s6-1. For the portrayal of a diagram showing relationship between the concepts assessable income, ordinary income, statutory income and exempt income.
Income Tax Assessment Act 1997 (Cth), s28-5. For a map of the Division highlighting the four alternative methods for claiming car expenses.
Income Tax Assessment Act 1997 (Cth), s100-15. For a diagrammatic overview of steps 1 and 2 in determining whether there is a taxable capital gain or capital loss.
Patents Act 1990 (Cth), s4. For a flowchart outlining the steps in getting and maintaining
a standard patent.
3 Turner, C., Australian Commercial Law (24th Ed, 2003) contains a foldout flowchart on The Law of Contract outlining the steps from formation to discharge of a contract. Woellner, R. H., Barkoczy, S., Murphy, S. and Evans, C., 2005 Australian Taxation Law, includes a diagrammatic overview of alternative appeal paths in challenging an income tax assessment, at p186 and a flowchart for The Simplified Tax System- A Flowchart, at p 944.
4 In Richard Walter Pty Ltd v Federal Commissioner of Taxation 95 ATC 4440, at 4444-4445, a diagram is included in the report to outline a complex trust arrangement relevant to the issue being considered by the Federal Court.
In Davis & Anor v Federal Commissioner of Taxation 89 ATC 4377, at p 4385, a diagram is employed to illustrate the facts of the case relating to as assignment of future income payments.
5 Income Tax Assessment Act 1997 (Cth), s4-10(3) and
s4-15(1).
half of a taxation law subject. It can also be used by legal practitioners,
the Federal Commissioner of Taxation and judges to formulate
and structure their
approach to taxation issues in legal proceedings. The skeleton of the flowchart
is reproduced as Figure 1.6
The flowchart can be transposed into electronic format, burned onto a CD- ROM
and/or accessed online. Each of the topic boxes that
form the building blocks
that make up the tree diagram or flowchart appear as screens online and each is
able to be hyperlinked to
the relevant statutory provisions, case authorities
and paragraphs in Australian income taxation law textbooks.
In electronic format it is not possible to represent 55cm x 85cm flowchart on
a single screen. However, it can be logically dissected
into six
“Roadmaps”, each capable of single screen representation. The 230
topic boxes are embodied in the roadmaps and
each is contained on an individual
screen.7 Four of the roadmaps are highlighted in bold in Figure 1.
The roadmap relating to Ordinary income is represented in Figure 2. Although
not
all the topic boxes have been reproduced in detail. An example of one of the
topic boxes is shown in Figure 3. The bold type
shows the hyperlinking to
statutory provisions, case law authorities and paragraphs in a taxation law
textbook.
How then do computer assisted flowcharts contribute to judicial reasoning
and analysis?
They portray the entire law, both statute and common or case law relevant to
the issue in a logical and structured format. The flowchart
provides an overview
of the particular area of law under consideration and shows the links and steps
to be followed in analysing
the legal problem. The flowchart also forms the
foundation for hyperlinking.
6 A full versions is available as, CCH Online Tax Flowchart Guide (2005), in hard copy.
7 The flowchart has been transposed into electronic format by CCH Australia Limited. It is not possible to represent the entire flowchart on one screen so the chart is conveniently dissected into six Roadmaps, each of which for the basis of Topic Screens which make up the 230 topic boxes. Each topic box is capable of being represented on an individual screen.
I understand there is a computer program capable of representing a large, say
A3, document on a single screen and allowing the reader
to zoom in on a
particular segment of the document or chart.
ii. hyperLinking
Hyperlinking is being used by judges to link words in their judgment to
precedent cases, statutory provisions and facts, which can
include images, both
still and moving.8 There is scope to make more effective use of this
technique by delivering judgments online or on CD-ROM or DVD and storing them on
these platforms.
In the context of teaching and learning, lecture notes and materials can be
uploaded onto the university’s intranet, for example
WebCT. The lecture
notes and materials can then be hyperlinked to relevant law sites, statutes or
cases.
Law textbooks now make use of hyperlinking through access to the
publisher’s web site. This can be done as a marketing technique
to
encourage adoption of the book as the required textbook for the subject. The
material on the web site linked to from the textbook
can include additional or
supplementary text and commentary, statutory provisions, cases or extracts from
cases, worked examples
and testing questions, with or without answers.9
An alternative is to have a CD-ROM shrink-wrapped to textbook, containing
structured materials hyperlinked to detailed materials.
How then does hyperlinking contribute to judicial reasoning and analysis?
Computers and information in electronic format are necessary
for hyperlinking.
The techniques assist by more effective searching for relevant authorities,
downloading the material and cutting
and pasting to analyse the material and
integrating it to structure the legal reasoning in coming to an opinion on the
issue. Judgments
themselves could be much shorter with much of the supporting
material, such as, statutory provisions, case law precedents and extracts
from
lower court judgments hyperlinked rather than reproduced in the
judgment.
iii. the LegaL paradigm
There is a well established legal paradigm employed by students,
practitioners and judges in learning and practicing law. Facts and
issues are
delineated, on the one hand, and the relevant law is collected together, on the
other. The law is then applied to the
facts to come to a decision or holding.
The law may
8 For example, see, Mobileworld Communications Pty Ltd v Q & Q Global Enterprise [2003] FCA (4 December 2003) at 8, which related to a trade mark image described as; MAN, CARTOON ATOP GLOBE.
9 See, for example, Gilders, F., Taylor, J., Richardson, G. and Walpole, M., Understanding
Taxation law: An Interactive Approach 2nd Ed
(2004).
embody precedents and as an outcome of the case, precedents may be created,
confined, extended or distinguished. The legal paradigm
can be represented in
flowchart format and each box in the chart, embodying facts or law can be
hyperlinked. By representing the
judgment in flowchart format the judicial
reasoning and analysis underlying the case is more apparent.
The legal paradigm is represented in diagrammatic format in Figure 4. It has
been applied to two leading taxation law cases. In The Commissioner of
Taxation v Whitfords Beach Pty Ltd,10 [Figure 5], the issue to be
decided was whether the profits from the sub-division and sale of land
constituted proceeds from the carrying
on a business or were attributable to the
mere realisation of an asset. According to the law (common law re-enforced by
statute),
on the one hand, isolated transactions can give rise to assessable
income and, on the other hand, the mere realisation of an asset
(according to
common law) does not give rise to assessable income. The High Court decided that
the sub- division and sale of land
did amount to carrying on a business and
extended the rule that the normal proceeds from carrying on a business
constitute assessable
income to state that an isolated transaction can
constitute carrying on a business. The diagrammatic representation of the case
makes
explicit the approach and structure and logic behind the Court’s
reasoning in coming to a decision and laying down a precedent.
The case of Sun Newspapers Ltd and Associated Newspapers Ltd v Federal
Commissioner of Taxation,11 [Figure 6], is a foundation case on
establishing a test for the distinction between losses and outgoings of a
capital or revenue nature.
The distinction is important because losses and
outgoings of a capital nature are not deductible where as those of a revenue
nature
are. The case is also notable because the “players” include
members of the Packer family and the test was formulated
by Dixon J. The
diagrammatic portrayal of the case is instructive in that it can be linked back
to previous common law capital v
revenue tests and projected forward to a line
of cases that have applied the “Dixon J” test.
How then do computers employed to graphically portray the legal paradigm
assist in contributing to judicial reasoning and analysis?
The technique shows
or demonstrates explicitly the steps, processes, links and logic in judicial
reasoning in coming to a decision
on a legal problem or issue and establishing
judicial precedents.
10 [1982] HCA 8; (1982) 150 CLR 355.
11 [1938] HCA 73; (1938) 61 CLR 337.
Analysis of Evidence
Scientific analysis of evidence has been attempted over the past 200 years.
Almost one hundred years separate the leading works by
Bentham on
evidence12 and Wigmore on proof.13 More recent work on the
subject has been by Professors William Twining, David Schum and Terence
Anderson.14
The primary task in these works is to analyse a mixed mass of evidence
available and classifying and placing each piece in its proper
place in the
scheme of proof and in making detailed inferences from stage to stage; finally
arriving at a conclusion upon the main
probandum or probanda.15
16
Flowcharts are an effective technique for undertaking the task with the
diagrammatic presentation of all the relations between all
the relevant evidence
and the ultimate probanda in a particular case. “The constituent elements
are simple propositions of
fact, each listed and numbered in a
‘key-list’ of evidence; the relations between the propositions are
depicted in the
chart by a system of symbols devised by the
author.”17 “The proposed method involves two steps:
analysis, which involves the preparation of a key-list of all the relevant
evidence
expressed as simple propositions; and synthesis, in the form of a chart
which depicts the relations between each item in the key-list
with all other
items.”18
“In this respect the chart method is analogous to the algorithm, that
is a precise set of instructions, capable of being presented
in diagrammatic
form, for solving a well-defined problem. By breaking down a complicated rule
(or body of data) into a number of
simple components and presenting each in turn
in a particular order, it can enable the reader to find his way around a complex
body
of rules and locate the answer to his particular
problem.”19
12 Bentham, Rationale of Judicial Evidence, Specially Applied to English Practice, 5 Volumes, Edited by John Stuart Mill (1827).
13 Wigmore. The Principles of Judicial Proof as given by Logic, Psychology and General
Experience and Illustrated in Judicial Trials.
14 See Twining, W, Theories of Evidence: Bentham and Wigmore (1985) 135.
15 “Evidence is always a relative term. It signifies a relation between two facts, the factum probandum, or proposition to be proved, and the factum probans, or material evidencing the proposition.... On each occasion the questions must be asked, What is the proposition (Probandum) desired to be proved? What is the Evidentiary Fact (Probans) offered to prove it?” Anderson, T and Twining, W, Analysis of Evidence: How to do Things with Facts (1991)
54.
16 See Twining, supra n 14 at 116, 125, 126.
17 Ibid, 126.
18 Ibid, 131.
19 Ibid, 133-4.
“The function of an algorithm is to present rules in a visually more
comprehensible form than conventional prose.”20 “They
can be used to organise into manageable form large quantities of data or other
material, for example evidence to be presented
at a lengthy trial or rules of
particular areas of a law such as property law, tax
law....”21
“[T]he chart method is a rather more flexible tool than Wigmore
suggests, since it may be used to chart other matters.... It
is theoretically
possible for the same person to apply the chart method to a case, from the point
of view of the actual arguments
put forward at the trial, or from the point of
view of an historian trying to analyse as detached as he can all the evidence
available
to him for whatever purpose he specifies, or even from the point of
view of a logician to reconstruct the explicit and implicit arguments
about
particular probanda....”22
The flowchart technique has not met with widespread success. However,
according to Twining, “like the algorithm the method [chart
method] seems
to offer considerable possibilities for use in connection with new information
technology. This is as yet a largely
unexplored field, but it seems quite
possible that Wigmore’s method will come into its own in the computer
age.”23
“Wigmore categorised the main relations between evidentiary
propositions in terms of assertion, denial, explanation and rival
assertion....
However, he did not provide a comparable vocabulary for differentiating the
principal ways in which evidentiary propositions
may be combined or accumulate
or otherwise be seen as allies tending to support or to negate the same
conclusion , either directly
or indirectly.”24 Twining
proposes five ways in which evidentiary propositions may be related, namely, by
conjunction, compound propositions, corroboration,
convergence or catenate
inferences (inference upon inference or a chain of
inferences).25
A hypothetical illustration of Wigmore’s method, incorporating
Twinings’ combining of propositions, is presented in Figure
7. To
establish guilt in a criminal case the prosecution may have to prove
propositions A and B and C. Defences against conviction
may be establishing Not
A or Not B or Not C or establishing D and/or E. Propositions F1 and F2 and F3
are compound in relation to
A1. G1 and G2 both independently corroborate
proposition B. H1, H2 and H3 are three independent items of circumstantial
evidence
which
20 Twining, supra n 15 at 433.
21 Ibid, 433-434.
22 Twining, supra n 14 at 134.
23 Ibid, 135.
24 Ibid, 180.
25 Ibid, 180.
converge to support proposition C. Finally, A1, A2 and A3 are a chain of
propositions that tend to establish A. On the other hand,
propositions I, J or K
will tend to disprove A+B+C. In addition either proposition L or M will be a
defence against conviction. Relations
between propositions on the defence side
can be more elaborate than shown in the illustration.
An example of an algorithm in taxation law is shown in Figure 8. It relates
to the general statutory provision for losses or outgoings
incurred in gaining
or producing assessable income.26
How then do computers employed to flowchart evidentiary propositions in
pursuit of proof in judicial proceedings or employed in constructing
algorithms
embodying legal rules assist in contributing to judicial reasoning and analysis?
In evidentiary analysis and synthesis
flowcharts, being a diagrammatic format,
are of great assistance in structuring and providing an overview of the inter-
relationships
between the various categories of propositions and the logic and
pathways towards proof of the ultimate proposition to be established.
Computers
are of assistance in constructing and manipulating the flowcharts by adding to
subtracting from or shifting the links between
propositions. They are also
necessary for hyperlinking to what lies behind the propositions. That may be
facts, rules, statutory
provisions or case law precedents, all of which go
together to explain the propositions and their inter-relationships. Using
computers,
it is also possible to have an overview flowchart on one screen which
then links to other sub-charts or screens that make up the
pieces in the
“jig-saw puzzle”.
26 Income Tax Assessment Act 1997 (Cth) Section 8-1 General Deductions
8-1(1) You can deduct from your assessable income any loss or outgoing to the extent that:
(a) it is incurred in gaining or producing your assessable income; or
(b) it is necessarily incurred in carrying on a business for the purpose of gaining or producing your assessable income.
8-1(2) However, you cannot deduct a loss or outgoing under this section to the extent that:
(a) it is a loss or outgoing of capital or of a capital nature; or
(b) it is a loss or outgoing of a private or domestic nature; or
(c) it is incurred in relation to gaining or producing your exempt income or your non- assessable non-exempt income; or
(d) a provision of this Act prevents you from deducting it.
In the case of algorithms, computers are more effective than manual
techniques in constructing the algorithm. More importantly in
electronic format
the relevant statutory provisions and/or case law extracts (facts, issues,
decisions or precedents) can be hyperlinked
to each step in the logical chain of
reasoning that underlies the algorithm.
Expert Systems
Expert systems are intended to provide “expert” problem-solving
advice in some area in which they are competent. Such
systems have been used in
medicine to diagnose diseases and to recommend suitable treatments. Although Law
is, in many ways, similar
to medicine expert systems have not been extensively
developed or applied in legal problem-solving. The widespread use of computers,
the significant development of computer programs and the availability of the
internet provide opportunities for developing and exploring
the application of
expert systems to legal problem- solving or judicial reasoning and
analysis.
With expert systems, the computer takes over from the human as the
“expert”. The human does not become redundant but is
integrally
involved in designing and programming the system and interacting with
it.
The attributes of an expert system, posed with a legal problem
include:
1. The system can ask questions of the lawyer. It can also ask successive
questions in proceeding through an algorithm.
2. The system has a perfect memory. It embodies a dynamic database with
links to all the materials in the database. It can call up
legal information the
lawyer has forgotten or is unaware of.
3. The system is flexible and non linear. Unlike a book, for example, where
material is presented in a linear format, a computerised
expert system can
package materials drawn from a variety of sources.
“What we are after in an expert system is a means not only of
representing the expert’s knowledge, but of systematically
extracting it
from the expert and refining
it.”27
27 Tyree, A, Expert Systems in Law (1989) 10.
Judicial reasoning and analysis involves both statute and case law. It is
relatively easy to build an expert legal system involving
statute law only.
Expert legal systems embodying case law or both statute and case law are more
difficult to build because judgments
involve a multitude of diverse
elements.
To build a computer based expert legal system the legal information needs to
be reduced to a computer program compatible format. The
information can take the
form of propositions relating to statutory provisions, facts, issues, decisions
and precedents. The system
then builds interrelationships between the
propositions. The system can then be represented in diagrammatic format, known
as a semantic
net or frame diagram.28 An example relating to
negotiable instruments is portrayed in Figure 9.29 The frame diagram
is based on the Bills of Exchange Act 1909 (Cth) and portrays the
creation, negotiation and discharge of a bill of exchange. It can be employed in
judicial reasoning an analysis
to deal with issues such as, the status of the
holder of the bill or the liability of various parties on the bill. Each box in
the
model can be hyperlinked to relevant sections in the Act and case law
authorities and likewise for each transaction between parties.
Another type of expert system that may find application in the Law is the
“Finite State Machine.” “Such a machine
is defined as follows:
there are a number of different “states;” at each state, the user
may be asked information and
the reply together with the value of other
variables of the system determines the next state that the machine
enters.”30 Tyree uses the law relating to negligence as an
example of a finite state machine.31 A variation of the model is
reproduced in Figure 10. The machine proceeds through the four elements that
need to be established in
order to find a party liable for negligence. It can be
used to structure legal reasoning and analysis and constructing judgments.
Once
again, each element in the machine can be hyperlinked to relevant principles and
precedents. The model can be employed as a
template for negligence cases with
judges simply plugging in the law relating to each element and arriving at a
decision as to liability.
Legal rules can be found in statutes and cases. In statutes they comprise
sections of an act, where as in cases they constitute the
ratio decidendi
of the case. In general, the rules take the form of; if... then. “The
difficulty of formulating rules which capture case law
reasoning may reflect the
fact that
28 See ibid, 43-46 for a more detailed explanation.
29 Ibid, 44 & 46, uses the law on negotiable instruments to provide an example of a semantic net and a frame diagram.
30 Ibid, 51.
31 Ibid, 54-57.
case law reasoning is closer to inductive than deductive reasoning. Although
reasoning with case law may have some deductive components,
the essence of it
would appear to be to generalise from a number of instances rather than the
application of logical rules.”32 One approach to formulating
such rules and applying them in judicial reasoning and analysis is to employ the
concept of “similarity”.
The approach involves collecting together all the relevant cases on a
particular legal subject. Each case contains a number of materially
relevant
facts and an outcome, such as guilty or not guilty or liable or not liable or,
more generally, win/lose. Each of the facts
can be stated as a proposition which
is either true or false and can be assigned a value “1” for true and
“0”
for false. Each case, on the particular subject, can then be
represented by a vector of “1s” and “0s”. The
range of
cases x the defining attributes form a matrix. The outcome of the case can be
added as a further column. A hypothetical
example is shown in Figure 11, with
rows of cases designated as C1 to C8 and the attributes or facts, in columns,
designated as F1
to F6. The attributes or facts F1-F6 form a vector for each
case, which leads to the outcome of W for win or L for lose. The next
step is to
assign weights to each of the facts, propositions or attributes. With the more
significant facts attracting higher weights.
Following this it is necessary to
measure the distance between the cases. The shorter the distance between the
cases the more likely
that the same outcome will apply on the basis of a
precedent common to those cases. The wider apart the distance between cases the
more likely there are distinguishing facts and outcomes between the cases and
the more likely a different precedent will apply. When
a new dispute arises,
judicial reasoning and analysis simply involves plugging the facts into the
existing matrix, doing the calculations
to see where the case in issue lies in
relation to established cases and applying the decision and precedent of the
closest cases.
A hypothetical example of a matrix showing distances between
cases is shown in Figure 12. It is up to the human expert to tease out
the
materially relevant facts and weight to assign to each one. The role of
computers is to employ statistical techniques for determining
the importance of
each attribute across the range of cases and to measure the distance between
cases. A worked example of this technique,
applied to finders’ cases, has
been demonstrated by Tyree.33
One difficulty in building and applying computer based expert legal systems
is that the information may need to be in quantitative
or numerical format.
Fortunately, as illustrated above, it may be sufficient to simply be able to
assign vales of “1”
or “0” to build expert legal systems
to be used in judicial reasoning and analysis.
32 Ibid, 133.
33 Ibid, 137-143.
Statistical Analysis of Interdependence
A judgement comprises a number of elements including, facts, issues, legal
precedents, statutory provisions and holdings, decisions
or opinions. Each of
these elements are variable, in the sense that they vary from case-to-case. One
role for the law is to collect
together similar variables from a line of cases
in order to establish principles of wider application to future cases. The
statistical
analysis of interdependence can assist in this regard, its goal is
to give meaning to a set of variables. The statistical analysis
techniques
include, factor analysis, multidimensional scaling and cluster analysis. The
techniques are dependant on computer programs
for their implementation and they
are capable of handling non-metric variables, of the type found in legal
proceedings.
The techniques have been widely employed in marketing for, for example,
positioning products in product space. They can also be applied
to position
political leaders in and along dimensions defining political space. It is
suggested that the techniques could be fruitfully
employed to position cases in
“legal space”. The dimensions in legal space can be made up of
vectors of facts derived
from decided cases. New cases can then be located in
legal space. Their position, in relation to clusters of decided cases will
provide
guidance as to the appropriate outcome and precedent to be
applied.
“Factor analysis is a multivariate statistical technique that addresses
itself to the study of interrelationships among a set
of observed variables....
The primary purpose is the resolution of a set of observed variables in terms of
new categories called
factors.”34 The factors can then be
rotated in order to more effectively describe the variables. Factor analysis is
useful in that it can:
1. Point out the latent factors or dimensions that determine the
relationship among a set of observed variables or vectors.
2. Point out relationships among observed variables that were there all the
time but not easy to see.
3. Be used when things need to be grouped or clustered.
In other words, “Factor analysis is basically a method for reducing a
set of data into a more compact form, while throwing certain
properties of the
data into bold relief. The user of factor analysis focuses on the set of
variables
34 William D. Wells and Jagdish N. Sheth, ‘Factor Analysis in
Marketing Research’ in Aaker, DA, Multivariate Analysis in Marketing:
Theory & Application (1971) 213.
for which information has been collected and poses the question: Can the
information contained in the original variables be summarised
in a smaller
number of new variables.”35
“The typical problem to be handled by the multi-dimensional-scaling
procedures might roughly be stated as follows: Given a set
of stimuli which vary
with respect to an unknown number of dimensions, determine (1) the minimum
dimensionality of the set and (2)
projections of the set of stimuli (scale
values) on each of the dimensions involved.”36
“The purpose of cluster analysis is to identify objects (or variables)
which are similar with respect to some criteria. The
resulting object clusters
should have high internal (within cluster) homogeneity and high external
(between clusters) heterogeneity).
Geometrically, the objects within a cluster
should be close together and the objects in different clusters should be far
apart.”37 As mentioned above factor analysis can be used to
cluster objects or variables. If a group of variables has in common high
loadings
on one factor (calculated by computer based statistical techniques),
then they are viewed as forming a cluster.38 “More formally
stated, the problem is: How should objects be assigned to groups so there will
be as much likeness within groups
and as much difference among groups as
possible.”39
A hypothetical example of how these statistically based multivariate
techniques might be employed to assist in judicial reasoning
and analysis,
without going into the statistical analysis, is as follows: Given a number of
cases on a particular legal issue or
topic, such as, negligent misstatement,
subsistence of copyright or tax avoidance, what materially relevant facts
determine that
some cases are similar to others in their outcome and what
materially relevant facts determine that groups of other cases are dissimilar
in
outcome. Further, what principle underlies each group of cases that are similar
to each other but differentiated from other groups.
It has been demonstrated, above, that the facts in cases can be
mathematically represented as vectors. It is now possible to represent
those
vectors in multi- dimensional legal space, which cannot be visualised but only
operated on by statistical techniques. However,
it is possible to provide a two
dimensional
35 William F. Massey, ‘What is a Factor Analysis? Research’ in Aaker, D. A., Multivariate
Analysis in Marketing: Theory & Application (1971) 241.
36 Warren S. Torgerson, Theory and Methods of Scaling (1958) 247-8.
37 Aaker, DA, Multivariate Analysis in Marketing: Theory & Application (1971) 299.
38 Ibid, 300.
39 Ronald E. Frank and Paul E. Green, ‘Numerical Taxonomy in Marketing Analysis: A Review
Article’ in Aaker, DA, Multivariate Analysis in Marketing: Theory & Application (1971)
303.
example, in Figure 13. Each of the vectors or variables represent facts in
legal space. Factor analysis will tease out a minimum number
of factors that
explain the variables. It is then up to the legal expert to label the factors.
Established cases, from which the
fact vectors have been derived are then
located in legal space. The legal expert can then spell out the legal principles
each clusters
have in common, with the assistance of the factors in the space.
This constitutes an exercise in judicial reasoning and analysis.
In the example,
C1+C2+C5 form one cluster, C3+C4+C6 form another cluster and C7+C8 form a third
cluster. Each cluster will have a
common underlying legal principle or rule. It
is the task of judicial reasoning and analysis to identify the common principle
or
precedent. Another task for the legal expert is to identify and label the
factors. If the legal issue related to tax avoidance, then
Factor 1 might be a
dimension designated as Tax Planning and Factor 2 might be a dimension
designated as Tax Avoidance.
Figure 1. Simplified Representation of the Income Tax Flowchart
Figure 2. Roadmap: Statutory Income
3.1 Income from personal exertion
[ 6-040, 7-000 – 7-020]
Ordinary income is interpreted as income according to the ordinary concepts
and usages of mankind [Scott v C of T (NSW)] and is categorized as
income from personal exertion or income from property. [97 s6(1)]
Income from personal exertion can be further sub-divided into income from
work or provision of services and proceeds from carrying
on a business.
Five categories of income from property are recognized, namely: annuities,
interest, rent, royalties and corporate distributions
(dividends).
A number of propositions have been developed by the courts in order to determine whether a receipt constitutes income from personal exertion or income from property.
Figure 3. Topic screen: Ordinary Income – Income from Personal
Exertion
Figure 4. The Legal Paradigm
Figure 7. Hypothetical Illustration of Relations Between
Evidentiary Propositions
Figure 8. Algorithm for Deductibility of Losses or Outgoings under
s8-1
Figure 9. Bills of Exchange Model: Creation, Negotiation and
Discharge
Figure 10. Negligence: Finite State Machine
Figure 11. Cases x Attributes and Outcomes Matrix
Figure 12. Measurement of the Distance between Cases
Figure 13. Fact Vectors and Location of Cases in Legal Space
NZLII:
Copyright Policy
|
Disclaimers
|
Privacy Policy
|
Feedback
URL: http://www.nzlii.org/nz/journals/NZYbkNZJur/2006/19.html