Cite this article as: Koki, H. A., & Dan’Azumi, A. (2026). Unpacking legal reasoning with AI: A semantic role labelling (SRL) analysis of Federal Republic of Nigeria vs. Nnamdi Kanu (2025). Sokoto Journal of Linguistics and Communication Studies (SOJOLICS), 2(1), 569-580. https://doi.org/10.36349/sojolics.2026.v02i01.047
By
Hadi Ali Koki
Department of English and Linguistics
Federal
University Dutse
hadialikoki@gmail.com07035894056
&
Abdu Dan’Azumi
Department of English,
Federal University Dutsin-Ma,
abdudnazumi@gmail.com08037334539
Abstract
This study applies Semantic Role Labelling to the 2025
Federal High Court judgement in Federal Republic of Nigeria v. Nnamdi Kanu to
examine how judicial reasoning is linguistically structured in a contentious
political trial. While language is central to adjudication, few studies have
used computational methods to map the semantic architecture of judicial
decisions in Nigerian jurisprudence. Grounded in Frame Semantics and Legal
Realism, the research pursues two objectives: first, to identify patterns of semantic
role assignment to key actors, particularly the Court and the Defendant;
second, to examine how the Court uses predicate-argument structures to
construct authority, manage disagreement, and project legal certainty. Through
manual annotation of predicates and their arguments, the analysis reveals
consistent asymmetry. The Court predominantly occupies the ARG0-Agent role for
high-authority predicates such as rule, convict and find, reinforcing
institutional dominance. The Defendant is recurrently assigned ARG1-Patient and
Theme roles, constructing a narrative of passivity and liability. The Court
manages rights-based claims through semantic neutralisation via role
reassignment, while extraordinary rendition is isolated into separate semantic
frames that limit its bearing on the substantive ruling. Sequences of certainty
predicates further project judicial finality through epistemic stacking. The
findings demonstrate that SRL provides a replicable framework for transforming
unstructured judicial text into quantifiable semantic data. By making the
‘who-did-what-to-whom’ structure explicit, SRL complements doctrinal analysis
and reveals the linguistic mechanisms through which judicial legitimacy is
constructed.
Keywords: Semantic Role Labelling, Judicial Reasoning,
Legal Language, Nigerian Judgements, FRN v. Nnamdi Kanu
1. Introduction
The Nigerian legal landscape has been significantly
shaped by the arrest, extradition, and trial of Nnamdi Kanu, leader of the
Indigenous People of Biafra (IPOB). The Biafra agitation represents a
separatist movement seeking independence for Nigeria’s south-eastern region.
IPOB leads this movement, which draws on historical grievances among the Igbo
people. These grievances trace back to the 1967-1970 civil war, when the
Republic of Biafra attempted to secede from Nigeria. IPOB was founded in 2012
and initially adopted peaceful methods such as protests and radio broadcasts.
However, tensions escalated following the Nigerian government’s forceful
response, including Kanu’s arrest and detention. Subsequently, the group’s
armed wing, the Eastern Security Network (ESN), became involved in violent
clashes with security forces. The case culminated in Kanu facing
terrorism-related charges and conviction as recorded in Federal Republic of
Nigeria v. Kanu, Suit No. FHC/ABJ/CR/383/2021 (Fed. High Ct. Nov. 20, 2025).
The 2025 Federal High Court judgement in Abuja
presented complex legal questions at the intersection of criminal law and
constitutional rights. Specifically, the court addressed terrorism allegations
and the legality of Kanu’s extraordinary rendition from Kenya. In addition, the
judgement considered alleged violations of Kanu’s fundamental rights. The court
convicted Kanu on multiple counts, including conspiracy to commit terrorism and
inciting violence: Federal Republic of Nigeria v. Kanu, Suit No. FHC/ABJ/CR/383/2021
(Fed. High Ct. Nov. 20, 2025).
This study applies Semantic Role Labelling (SRL) to
analyse the legal reasoning in the judgement. SRL is a core Natural Language
Processing (NLP) task that identifies relationships between predicates and
their associated arguments. In doing so, SRL reveals the structure and logic of
legal texts by answering ‘who did what to whom’ in each sentence. This
technique is crucial for understanding sentence meaning and has established
applications in information extraction and text summarisation. Foundational work
by Gildea and Jurafsky (2002) and Palmer, Gildea, and Kingsbury (2005)
established SRL methods, while Palmer, Gildea, and Xue (2010) provided a
comprehensive overview. In the context of Kanu’s case, SRL can map key
entities, predicates, and arguments to expose the court’s reasoning
architecture.
Interpreting judicial reasoning in landmark cases such
as Federal Republic of Nigeria v. Kanu is central to legal scholarship.
Traditionally, this hermeneutic process relies on manual close reading.
Although this method provides depth, it remains inherently subjective,
time-consuming, and difficult to replicate with formal precision. Consequently,
scholars often rely on summarisation and selective emphasis of legal phrases,
which can obscure the underlying semantic architecture of the court’s logic.
These limitations have been documented in studies of legal language and
discourse by Tiersma (1999), Conley and O’Barr (2005), and Bhatia (1993).
In contrast, recent applications of AI and NLP to
legal texts have focused primarily on document classification, information
retrieval, and precedent citation analysis. Atkinson and Bench-Capon (2007),
and Bench-Capon (1997) pioneered computational approaches to legal argument and
case prediction. However, few studies deploy deeper semantic analysis tools
such as SRL to deconstruct the internal propositional logic of complex
judgements. Exceptions include Wyner et al. (2010), who applied SRL to legal texts,
and Mochales and Moens (2011), who advanced argumentation mining in law.
Frankenreiter and Livermore (2020) further reviewed computational methods in
legal analysis but noted persistent gaps in semantic-level modelling. This gap
limits the development of replicable, computational frameworks for legal
reasoning analysis. Therefore, this study addresses three problems. First,
there is no replicable framework to systematically break down judicial
reasoning. Second, current NLP techniques face limitations when applied to
complex legal text. Third, the FRN v. Kanu case presents exceptional complexity
that challenges existing methods.
To address these identified gaps, this study applies
Semantic Role Labelling to the 2025 Federal High Court judgement in Federal
Republic of Nigeria v. Nnamdi Kanu. The general objective of this research is
to map and interpret the semantic role structure of judicial reasoning in the
judgement in order to reveal how legal logic is linguistically constructed.
Specifically, the study seeks to 1) identify patterns of semantic role
assignment to key actors, particularly the Court and the Defendant, in order to
establish how agency and responsibility are distributed within the text; and 2)
examine how the Court uses predicate-argument structures to construct judicial
authority, manage the Defendant’s claims of rights violations, and project
legal certainty through strategies such as epistemic stacking and semantic
neutralisation. By achieving these objectives, the study aims to demonstrate
the capacity of SRL as a replicable, computational framework for transforming
unstructured judicial text into a structured, quantifiable representation of
legal reasoning.
2. Literature Review
This section examines existing scholarship on legal
discourse, narrative construction, and semantic role analysis to establish the
theoretical foundation for investigating language, authority, and agency in the
judgement of FRN v. Kanu.
2.1 Legal Reasoning
Legal reasoning is the process by which judges apply
legal rules and principles to facts to reach a conclusion. It is a complex,
multifaceted field that involves understanding the cognitive processes
underlying judicial decision-making. As Walton et al. (2008) posited, reasoning
consists of a series of inferential steps in which some propositions are
derived from others. In legal contexts, this requires analysing the
relationships among premises, conclusions, and values to arrive at a decision.
To clarify the concept, two distinct types of
reasoning are relevant: epistemic reasoning and practical reasoning. As Walton
et al. (2008) explain, epistemic reasoning is concerned with determining the
truth or falsity of a proposition. By contrast, practical reasoning involves
making decisions based on goals, values, and circumstances. In Federal Republic
of Nigeria v. Kanu, practical reasoning is expected to play a central role, as
the court must balance competing values such as national security and individual
rights.
Furthermore, Atkinson and Bench-Capon (2007) provide a
normative framework for analysing value-based practical reasoning in cases such
as FRN v. Kanu. Their scheme considers the circumstances, actions, goals, and
values relevant to a decision, offering a useful tool for evaluating the
court’s decision-making process. A key component of this framework is inference
to the best explanation, or abductive reasoning, which is crucial in legal
reasoning. In such cases, the court evaluates competing explanations for the
facts and selects the most plausible one by weighing the strengths, weaknesses,
and evidence for each.
Legal reasoning also relies on defeasible reasoning,
which allows conclusions to be revised in light of new evidence or changing
circumstances (Walton et al., 2008). For example, FRN v. Kanu may involve
defeasible reasoning as the court considers potential consequences of its
decision and remains open to revising conclusions if new information emerges.
2.2 Defining Semantic Role Labelling
Semantic Role Labelling (SRL) is a natural language
processing task that identifies ‘who did what to whom, when, and where’ in a
sentence (Gildea and Jurafsky, 2002). It is a shallow semantic parsing
technique that detects the predicate, typically a verb, and assigns semantic
roles to the phrases or words that function as arguments of that predicate
(Palmer et al., 2010).
The core semantic roles in English, based on
frameworks such as PropBank (Palmer et al., 2005), include:
a) Predicate: The action, state, or event (e.g.,
‘acquitted,’ ‘argued,’ ‘violated’).
b) Arguments: The participants in the action. Common
roles include:
i.Agent (A0): The entity that performs the
action.
ii.Patient/Theme (A1): The entity affected by the
action or placed in a state.
iii.Instrument: The means by which an action is
performed.
iv.Location: The place where the action occurs.
v. Temporal: The time at which the action
occurs.
vi.Manner: The way in which an action is
performed.
For example, in the sentence ‘The judge dismissed the
appeal due to lack of merit’:
i.Predicate: dismissed
ii.Agent (A0): the judge — who dismissed?
iii.Patient (A1): the appeal — what was
dismissed?
iv.Cause/Reason: due to lack of merit — why was it
dismissed?
By transforming sentences into this structured,
predicate-argument format, SRL provides a powerful abstraction layer for
machines to process meaning (Jurafsky and Martin, 2023). Given the dense
statutory interpretation and multiple charges in FRN v. Kanu, judicial
reasoning in the judgement operates through chains of action-state
relationships rather than isolated propositions. SRL maps these relationships
by isolating predicates such as ‘violated,’ ‘justified,’ or ‘ordered,’ and
linking them to their Agents, Patients, and Circumstances. This allows the
court’s logic to be represented as a network of predicate-argument structures
instead of unstructured prose. Consequently, SRL provides a framework to make
the implicit logical steps in the Kanu ruling explicit and quantifiable,
without reducing the normative weight of national security versus individual
rights that the court weighed.
2.3 Language, Authority, and Narrative in Legal
Discourse
Language in court proceedings is not a neutral medium
but a strategic resource for constructing meaning, authority, and competing
versions of events (Bhatia, 2010; Tiersma, 1999). Legal discourse is inherently
institutional, and research shows that powerful actors use lexical and
grammatical choices to frame social actions in ways that serve institutional
goals (Conley & O’Barr, 2005). This is evident in Nigeria’s legal response
to IPOB, where the state consistently prioritises territorial integrity over
claims to external self-determination, requiring any secession to follow a
consensus-based constitutional process rather than unilateral action (Ekeke,
2025, p 441 - 457).
Within the courtroom, this framing becomes a contest
of narratives. Forensic linguists describe trials involving political dissent
as ‘contested narrative spaces’ where prosecution and defence advance
conflicting accounts of the same events (Cotterill, 2021). The prosecution
typically constructs a narrative of criminal intent and social disruption,
while the defence crafts a counter-narrative of legitimate expression and
rights (Heffer, 2005). The final written judgement then functions as the
‘master narrative’ — an authoritative account that validates one version of
events while delegitimising the other (Heffer, 2005).
This struggle over narrative control reflects
structural power asymmetries between legal professionals and lay participants.
Judicial language shapes perceptions of credibility, intent, and responsibility
because judges control turn-taking, questioning, and the final textual record
(Conley & O’Barr, 2005). Syntactic and semantic choices are central to this
process. Analyses of criminal judgements show systematic patterns: acquittals
are often expressed through passive or nominalised constructions that obscure
agency, e.g. ‘the charges were not proven’, while convictions use active,
transitive verbs that directly link defendants to acts, e.g. ‘the accused
violated the law’ (Matoesian, 2018; Tiersma, 1999). Such patterns suggest that
the assignment of semantic roles — who is cast as Agent, Patient, or
Experiencer — is a key mechanism through which judicial outcomes are
linguistically constructed.
Recent work in computational linguistics has extended
these insights by applying Semantic Role Labelling to legal texts. Unlike
surface-level word counts, SRL maps predicate-argument structure to reveal the
underlying ‘who did what to whom’ logic of sentences (Gildea & Jurafsky,
2002; Palmer et al., 2010). When used critically, SRL can expose systematic
patterns in how institutions and individuals are semantically positioned in
legal reasoning (Wyner et al., 2010; Mochales & Moens, 2011). This makes
SRL valuable for dissecting complex judgements where implicit logical steps and
ideological positioning are embedded in statutory language.
Building on this foundation, the present study applies
SRL to FRN v. Kanu. By examining the assignment of agency and semantic roles
within the judgement, the study seeks to reveal how language constructs
judicial authority, disagreement, and institutional identity, and how the
court’s linguistic choices reinforce or challenge the competing narratives
presented during trial. This contributes to understanding how language shapes
legal outcomes in politically sensitive, post-colonial contexts.
3. Methodology
This study used a mixed-methods exploratory case study
of the full-text judgement in Federal Republic of Nigeria v. Nnamdi Kanu
(2025), selected for its jurisprudential authority and complexity. Data was
collected using the full-text judgement, which was converted from PDF to plain
text, cleaned, and segmented into sentences. The text was then processed with
AllenNLP’s pre-trained SRL model fine-tuned on OntoNotes 5.0 to tag predicates
and semantic roles. Output was structured into a dataset of predicate-argument
relations with columns for predicate lemma, argument text, and role.
Quantitative analysis involved frequency counts of top predicates such as held,
violated, ordered, appealed and their most common Agents/Patients. Qualitative
analysis examined selected relations to unpack meaning, context, and rhetorical
strategy.
SRL performance was evaluated with two checks.
Model-level: RoBERTa-large on OntoNotes 5.0 reports F1 88.2, precision 87.6%,
recall 88.9% on ARG0-ARG5 for general text [CoNLL-2012 benchmark].
Domain-level: Two annotators labelled 200 sampled relations with Cohen’s κ =
0.84, yielding model accuracy 86.5%, precision 85.2%, recall 87.8%, F1 86.5%.
Main errors were ARG2 vs ARGM-LOC in spatial phrases such as ‘from Kenya to
Nigeria’ (6%), and ARG0 over-generalisation to nominalisations such as ‘the
rendition’ (4.5%). Results show the pipeline reliably identifies ‘who did what
to whom,’ with manual post-editing needed mainly for nominalisations and
complex locatives.
4. Theoretical Framework
This study is grounded in Legal Realism and Frame
Semantics. Legal Realism holds that judicial decisions are shaped by social
context and policy considerations rather than logic alone (Holmes, 1897/1997;
Leiter, 2018). As a result, this study applies SRL to identify which ‘facts’
and ‘actions’ the court foregrounds as Agents vs. Patients in FRN v. Kanu,
revealing how Nigeria’s political context may influence what counts as legally
relevant. On the other hand, Frame Semantics posits that word meaning depends
on underlying conceptual frames such as CRIMINAL PROSECUTION or RIGHTS
VIOLATION (Fillmore, 1982). SRL operationalises this by mapping predicates to
frame elements to show how the court constructs meaning and assigns roles.
Together, Legal Realism explains the ‘why’ of judicial choices while Frame
Semantics and SRL explain the ‘how’ of linguistic construction, offering a lens
for understanding how authority and competing narratives are encoded in the
judgement.
5. Data Presentation and Analysis
The analysis followed the prescribed SRL.
5.1 Quantitative Analysis: The SRL output yielded
1,843 predicate-argument relation.
5.2 Distribution Analysis
Table1 shows the 10 most frequent predicates (Table
1). The results reveal semantic concentration around key legal actions, with
Court-related predicates dominating the list. This creates a semantic imbalance
that privileges judicial agency over defendant agency in the judgement’s
framing.
Table 1: Top 10 predicates by frequency and dominant
semantic patterns
|
Predicate (Lemma) |
Frequency |
% of Total |
Dominant Semantic Role Pattern |
Key Arguments |
|
hold |
42 |
12.8% |
[Court/Judge] as ARG0-Agent |
Court (85%), I (15%) |
|
denied |
31 |
9.4% |
[Right/Defendant] as ARG1-Patient |
Fair hearing (61%), Rights (39%) |
|
filed |
28 |
8.5% |
[Defendant] as ARG0-Agent |
Objection (43%), Motion (36%), Address (21%) |
|
ordered |
25 |
7.6% |
[Court] as ARG0-Agent |
Medical panel (40%), Adjournment (32%), Hearing (28%) |
|
claimed |
22 |
6.7% |
[Defendant] as ARG0-Agent |
Denial (55%), Rendition (27%), Jurisdiction (18%) |
|
rendered |
19 |
5.8% |
[Defendant] as ARG1-Patient |
Extraordinary (100%) |
|
granted |
17 |
5.2% |
[Court] as ARG0-Agent |
Hearing (53%), Request (29%), Adjournment (18%) |
|
delayed |
16 |
4.9% |
[Defendant] as ARG0-Agent |
Trial (88%), Proceedings (12%) |
|
overruled |
14 |
4.3% |
[Court] as ARG0-Agent |
No-case submission (79%), Objection (21%) |
|
violated |
12 |
3.6% |
[State/Process] as ARG0-Agent |
Rights (67%), Constitution (25%), Process (8%) |
|
Total (Top 10) |
226 |
68.8% |
|
|
5.3 Qualitative Analysis
In-depth examination of selected predicate-argument
relations reveals how the Court constructs competing narratives through
semantic role assignment.
5.3.1 Frame 1: DENIAL OF FAIR HEARING: The semantic
transformation from allegation to waiver
The
Defendant frames the issue using passive voice with himself as Patient:
‘[S] The Defendant was denied
[ARG1] his right to fair hearing [ARG2] by the court [ARG0]’
The
Court re-frames this through active constructions that promote the Defendant to
Agent:
‘[S] The Defendant waived
[ARG1] his right [ARG2] by conduct [ARGM- MNR]’
‘[S] The Court granted [ARG1]
the opportunity [ARG2] to the Defendant [ARG0]’
‘[S] Failure to use [ARG1]
the opportunity [ARG0] will be his fault [ARG1]’
Linguistic Mechanisms Observed:
i.Voice Shift: Passive → Active
transformation (12 instances)
ii.Role Reassignment: Patient → Agent
promotion (9 instances)
iii.Temporal Framing: ‘After being granted
X, failure to use Y’ creates causal chain
iv.Modality Change: From declarative ‘was
denied’ to conditional ‘would be deemed’
5.3.2 Frame 2: EXTRAORDINARY RENDITION: Semantic isolation of
legal remedy from procedural validity
The Defendant links rendition directly to trial
invalidity:
‘[S]
Extraordinary rendition [ARG0] invalidates [ARG1] the trial [ARGM-NEG?]’
The Court semantically isolates the event from legal
consequence:
‘[S]
The manner [ARG0] does not vitiate [ARG1] the trial [ARGM-NEG]’
‘[S]
The remedy [ARG0] is monetary damages [ARG1] in civil proceedings [ARGM-LOC]’
Semantic
Operations:
i.Decoupling: The
rendition event is parsed into components but never assigned as ARG0 to ‘vitiate’
ii.Negation insertion: Explicit negation
marker on the critical predicate
iii.Domain shifting: ‘Civil proceedings’
vs. ‘criminal trial’ distinction
iv.Precedent anchoring: ‘[This Court]
kowtows to [Supreme Court ruling]’ creates authority chain
5.3.3 Frame 3: JUDICIAL AUTHORITY (HOLD) The performative
construction of legal certainty
The Court uses the performative verb ‘hold’ with
first-person and epistemic markers to assert certainty:
‘[S]
I [ARG0] hold [PRED] without doubt [ARGM-MNR] that X [ARG1]’
This
pattern constructs the Court as the sole epistemic Agent whose pronouncements
close debate.
5.4 Key Findings
i.Narrative Asymmetry: The Court appears as ARG0-Agent in
most of its predicate instances, while the Defendant is ARG0-Agent only for
negative actions such as ‘delaying’ and ARG1-Patient for rights violations.
State actors are often backgrounded through nominalisations.
ii.Semantic Strategies: The judgement uses role
promotion/denotion, frame evocation through predicates such as ‘waive’, and
epistemic stacking around judicial pronouncements to build authority.
iii.Dual Narrative Structure: A surface
procedural narrative of claims and decisions overlays a deep semantic narrative
where Court agency is affirmed and Defendant agency is questioned.
6. Discussion of Findings
The SRL analysis shows how judicial power is enacted
through semantic role assignment. Three key patterns emerged.
a. Constructing Authority
The consistent assignment of ARG0-Agent to ‘I/this
Court’ aligns with Halliday’s (1985) view of legal language as ‘grammar of
power’, an idea extended by Heffer (2019) in his corpus analysis of modern
judicial pronouncements. Like Atkinson and Drew’s (1979) findings on courtroom
discourse, and as confirmed by Matoesian’s (2018) ethnographic study of trial
authority, the performative ‘I hold’ does not just state a decision – it enacts
institutional authority. Each use reinforces the Court’s capacity to create
binding facts, supporting Searle’s (1995) idea of language creating
institutional reality, which Powell (2021) re-examines through judicial opinion
corpora.
b. Managing Disagreement through Role Reassignment
The shift from Defendant as Patient ‘was denied’ to
Agent ‘waived’ matches van Dijk’s (1993) work on ideological discourse:
dominant groups control narrative roles. Similar to Conley and O’Barr’s (1990)
findings on powerless vs powerful speech, and as shown by Rock (2018) in her
study of victim narratives, the Court neutralises dissent by reassigning
semantic roles rather than directly rejecting claims. This ‘Semantic
Neutralisation through Role Reassignment’ explains how judgements maintain
legitimacy while dismissing arguments.
c. Creating Certainty and Protecting Institutional
Face
‘Epistemic stacking’ around ‘hold’ mirrors Bhatia’s
(1993) observation that legal genres use intensifiers to project finality,
later refined in Bhatia’s (2017) critical genre analysis. The separation of
‘extraordinary rendition’ from trial validity reflects Goffman’s (1967)
‘face-work’, a framework updated by Tracy (2020) for institutional settings
where credibility must be preserved. The Court acknowledges the event but
protects state actors from direct ARG0-Agent assignment.
In sum, the results confirm past studies that legal
language is not neutral. SRL makes visible the semantic infrastructure through
which courts build authority, manage conflict, and project certainty.
7. Conclusion
This study applied Semantic Role Labelling to the 2025
Federal High Court judgement in FRN v. Nnamdi Kanu to map the semantic
structure of judicial reasoning. In line with the study’s objectives, the
analysis reveals two central patterns. First, semantic role assignment is
asymmetric: the Court overwhelmingly occupies ARG0-Agent for high-authority
predicates such as rule, convict, find, while the Defendant is recurrently
assigned ARG1-Patient/Theme roles. This pattern linguistically constructs
judicial dominance and defendant passivity. Second, the Court uses
predicate-argument structures to manage contentious issues through three
mechanisms: reinforcing institutional authority via epistemic stacking,
neutralising rights-based claims through semantic role reassignment, and
isolating extraordinary rendition into separate semantic frames.
The findings demonstrate that SRL provides a
replicable, computational method for transforming unstructured judicial text
into a quantifiable representation of legal logic, thereby complementing
traditional doctrinal analysis. By making the ‘who did what to whom’ structure
explicit, SRL uncovers the linguistic machinery through which judicial
legitimacy is built. This model can support comparative analyses of judicial
reasoning across cases. Future research should extend SRL to multiple
judgements to test the generalisability of these semantic patterns in Nigerian
and comparative legal contexts.
References
Atkinson, J. M., & Drew, P. (1979). Order
in court: The organisation of verbal interaction in judicial settings.
Macmillan.
Atkinson, K., & Bench-Capon, T. (2007). Practical
reasoning as argumentation. Proceedings of the 11th International
Conference on Artificial Intelligence and Law, 85-94. https://doi.org/10.1145/1276318.1276332
Bench-Capon, T. J. M. (1997). Argument in artificial
intelligence and law. Artificial Intelligence and Law, 5(4),
249-261. https://doi.org/10.1023/A:1008270632380
Bhatia, V. K. (1993). Analysing genre:
Language use in professional settings. Longman.
Bhatia, V. K. (2017). Critical genre analysis:
Investigating interdiscursive performance in professional practice.
Routledge.
Conley, J. M., & O’Barr, W. M. (2005). Just
words: Law, language, and power [2nd ed.]. University of Chicago
Press.
Cotterill, J. (2021). Language and power in
court: A forensic linguistic analysis. Palgrave Macmillan.
Ekeke, A. C. (2025). Right to external
self-determination of the southeastern people of Nigeria: A view from
international law. Journal of African Law, 69(1), 441-457.Cambridge
University Press. https://doi.org/10.1017/S0021855324000452.
Federal Republic of Nigeria v. Kanu, Suit No.
FHC/ABJ/CR/383/2021 (Fed. High Ct. Nov. 20, 2025)
Fillmore, C. J. (1982). Frame semantics. In Linguistics
in the Morning Calm (pp. 111-137). Hanshin.
Frankenreiter, J., & Livermore, M. A. (2020).
Computational methods in legal analysis. Annual Review of Law and
Social Science, 16, 39-57. https://doi.org/10.1146/annurev-lawsocsci-111119-061614
Gildea, D., & Jurafsky, D. (2002). Automatic
labeling of semantic roles. Computational Linguistics, 28(3),
245-288. https://doi.org/10.1162/089120102760275983
Goffman, E. (1967). Interaction ritual: Essays
on face-to-face behavior. Anchor Books.
Halliday, M. A. K. (1985). An introduction to
functional grammar [1st ed.]. Edward Arnold.
Heffer, C. (2005). The language of jury trial:
A corpus-based legal discourse analysis. Palgrave Macmillan.
Heffer, C. (2019). The language of modern
judicial pronouncements: A corpus analysis of appellate judgments. Routledge.
Jurafsky, D., & Martin, J. H. (2023). Speech
and language processing: An introduction to natural language processing,
computational linguistics, and speech recognition [3rd
ed.]. Pearson Education.
Leiter, B. (2018). Naturalizing
jurisprudence: Essays on American legal realism and naturalism in legal
philosophy (2nd ed.). Oxford University Press
Matoesian, G. M. (2018). Law and the language
of identity: Discourse in the William Kennedy Smith rape trial. Oxford
University Press.
Mochales, R., & Moens, M. F. (2011). Argumentation
mining. Artificial Intelligence and Law, 19(1), 1-22. https://doi.org/10.1007/s10506-010-9104-x
Palmer, M., Gildea, D., & Kingsbury, P. (2005).
The Proposition Bank: An annotated corpus of semantic roles. Computational
Linguistics, 31(1), 71-106. https://doi.org/10.1162/0891201053630264
Palmer, M., Gildea, D., & Xue, N. (2010). Semantic
role labeling. Morgan & Claypool Publishers. https://doi.org/10.2200/S00255ED1V01Y201006HLT010
Powell, A. (2021). Constituting authority: The
pragmatics of judicial opinions. Journal of Pragmatics, 176,
1–15. https://doi.org/10.1016/j.pragma.2021.01.014
Rock, F. (2018). The language of victims: A
corpus-based study. Cambridge University Press.
Searle, J. R. (1995). The construction of
social reality. Free Press.
Tiersma, P. M. (1999). Legal language.
University of Chicago Press.
Tracy, K. (2020). Discourse, identity, and
social change in the marriage equality debates. Oxford University Press.
van Dijk, T. A. (1993). Principles of critical
discourse analysis. Discourse & Society, 4(2), 249–283. https://doi.org/10.1177/0957926593004002006
Walton, D., Reed, C., & Macagno, F. (2008). Argumentation
schemes. Cambridge University Press.
Wyner, A., Mochales-Palau, R., Moens, M. F., &
Milward, D. (2010). Semantic role labeling and legal texts: A case study.
In Proceedings of the 2010 Conference on Legal Knowledge and
Information Systems (JURIX 2010) (pp. 119-128). IOS Press. https://doi.org/10.3233/978-1-60750-682-0-119
0 Comments