1. Introduction
Conceptual Metaphor Theory (CMT) posits that abstract concepts (target domains) are primarily understood through information mapped from more concrete, experiential source domains (Lakoff, Reference Lakoff and Ortony1993; Lakoff & Johnson, Reference Lakoff and Johnson1980). However, a standing theoretical preoccupation within the development of CMT has been explaining the systematic patterns of variation observed across different languages and cultures (Gibbs, Reference Gibbs2026; Kimmel, Reference Kimmel2004; Kövecses, Reference Kövecses2012; Yu, Reference Yu1995). This raises a fundamental question about the cognitive mechanisms that drive this variation. Kövecses (Reference Kövecses2015) and Abdel-Raheem (Reference Abdel-Raheem2020) have demonstrated that both immediate and broader cultural contexts are fundamental to where metaphors come from, how they vary, and how they are used. While extant scholarship has extensively examined the role of the individual (e.g. Ahrens, Reference Ahrens, Perrez, Reuchamps and Thibodeau2019; Cameron, Reference Cameron2007; Cameron & Low, Reference Cameron and Low2004), physical embodiment (e.g. Gibbs, Reference Gibbs2006; Gibbs et al., Reference Gibbs, Lima and Francozo2004; Yu, Reference Yu and Gibbs2008) and history (Burgers & Ahrens, Reference Burgers and Ahrens2020; Tay, Reference Tay2017; Zeng et al., Reference Zeng, Burgers and Ahrens2021) as determinants of metaphorical variation, far less is known about how group-level social identities shape metaphor choice, particularly within complex multifactorial political settings. This study addresses this gap by investigating the extent to which ethnicity, region and religion correlate with the metaphorical conceptualisation of NATIONHOOD within Nigerian presidential rhetoric.
Nigeria, formed through the amalgamation of over 250 distinct ethnolinguistic groups (Sircar, Reference Sircar1968), presents a rich context for investigating this question. The country’s sociopolitical landscape has historically revolved around three major identity pillars: ethnicity, region and religion (Owen & Usman, Reference Owen and Usman2015), which have cultivated distinct communities of experience. While the political implications of these identity markers are discussed in the literature (see Isiaq et al., Reference Isiaq, Adebiyi and Bakare2018; Opone & Omosor, Reference Opone and Omosor2025), their capacity to function as cognitive filters warrants scrutiny. In a deeply pluralistic society, these affiliations do not merely dictate political allegiance; they fundamentally shape the conceptualisation of the state itself. Therefore, Nigerian political discourse serves as an ideal context for observing how macro-social identities manifest linguistically.
To capture the complexity of these cognitive inputs, this study advances a multifactorial model that moves beyond traditional single-variable analyses. We posit that a leader’s ethnolinguistic identity, more than their regional or religious affiliation, serves as the primary predictor of their metaphorical conceptualisation of NATIONHOOD. Drawing on the Corpus of Nigerian Presidential Speeches (CORNIPS) and employing log-linear modelling, this paper demonstrates that Nigerian leaders from diverse backgrounds employ distinct metaphorical patterns when framing NATIONHOOD. These linguistic choices are not arbitrary; rather, they are closely linked to the specific historical and political experiences of the leaders’ respective groups, suggesting that different social realities engender distinct dominant cognitive models for the same abstract concept. The next section provides a review of the relevant literature.
2. Research background
2.1. Extralinguistic factors and metaphor use
The selection of region, ethnicity and religion as the primary extralinguistic variables for this study constitutes a principled decision grounded in the sociopolitical realities of Nigeria. As a highly pluralistic society (Chidozie & Orji, Reference Chidozie and Orji2024; Kur et al., Reference Kur, Agudosy and Orhewere2015), Nigerian political discourse and systemic stability have historically been contingent upon the delicate management of balance, representation and power rotation across these three identity pillars (Owen & Usman, Reference Owen and Usman2015; Tuki, Reference Tuki2024). The political behaviour of both the elite and the electorate is strongly shaped by ethno-regional and religious fault lines (Isiaq et al., Reference Isiaq, Adebiyi and Bakare2018), which, in many cases, have been shown to dictate voting patterns during general elections (Nwankwo, Reference Nwankwo2019; Ostien, Reference Ostien2012). Despite Nigeria operating a nominally federal system, political power remains heavily centralised. Consequently, intense competition among ethnic blocs for executive control has emerged as a defining feature of the nation’s political ecology.
Furthermore, the enduring struggle between national unity and sectional identity is continually negotiated within political rhetoric. For instance, in his 2015 inaugural address, President Muhammadu Buhari unequivocally stated, ‘I belong to everybody, and I belong to nobody’. This paradoxical statement may be interpreted as a strategic attempt to project neutrality and transcend the deep-seated pluralism of the polity he was elected to govern. However, such official rhetoric of unification often belies the realities of political praxis, as was the case with his administration. Recent scholarship (e.g. Opone & Omosor, Reference Opone and Omosor2025) establishes that Nigerian leadership, including the Buhari and Tinubu administrations, continues to operate along deeply entrenched ethnic, religious and regional lines. Such ethnification or nepotism is often evident in political appointments, where patronage frequently aligns with shared identity markers (Agbiboa & Maiangwa, Reference Agbiboa and Maiangwa2013). Essentially, these cleavages are not confined to elite politics but are also reflected in the broader societal discourse extending widely into digital and social media spheres (Aboh et al., Reference Aboh, Ganaah and Chigbu2025; Chigbu et al., Reference Chigbu, Aboh and Ganaah2025). Ultimately, the pervasive nature of these divisions renders ethnicity, region and religion indispensable variables for analysing Nigerian political rhetoric, particularly concerning the metaphorical conceptualisation of NATIONHOOD.
While the study of metaphor variation is well-established (e.g. Charteris-Black, Reference Charteris-Black, Musolff and Zinken2009; Chigbu & Ahrens, Reference Chigbu and Ahrens2025; Kövecses, Reference Kövecses2012; Sun et al., Reference Sun, Kong and Zhou2023; Zeng et al., Reference Zeng, Tay and Ahrens2020), and the role of context in shaping metaphorical production is widely acknowledged (Abdel-Raheem, Reference Abdel-Raheem2020; Kövecses, Reference Kövecses2015), much of the existing scholarship remains monofactorial, isolating the influence of a single variable, such as gender, political ideology or speaker role on metaphor use (e.g. Ahrens, Reference Ahrens, Perrez, Reuchamps and Thibodeau2019). The few multifactorial studies conducted have primarily focused on intra-linguistic or usage-based features (e.g. Garlepow et al., Reference Garlepow, Funke and Güldenring2025; Nicaise, Reference Nicaise2010; Tay, Reference Tay2016; Wu & Liu, Reference Wu and Liu2025). A multifactorial approach is therefore imperative to ascertain whether these macro-social identity variables pattern independently or interactively in structuring metaphorical discourse.
2.2. Studies on metaphors of NATIONHOOD
Studies in critical discourse and cognitive linguistics have increasingly recognised the role of metaphor in the discursive construction of nationhood. Far from being mere rhetorical ornaments, metaphors are understood as cognitive tools that structure our understanding of abstract political concepts (see Brugman et al., Reference Brugman, Burgers and Vis2019), making them tangible (Lakoff & Johnson, Reference Lakoff and Johnson1980) and emotionally resonant. As Šarić and Stanojević (Reference Šarić, Stanojević, Šarić and Stanojević2019) contend, the very idea of a nation requires constant ‘discursive imagination and invention’, a process in which metaphor is central to defining national values, borders and identity. A review of the literature reveals several dominant and recurring metaphorical frameworks used to conceptualise the nation, while also revealing culture-specific patterns.
The most prevalent of these conceptual frames is the personification of the nation. Extensive work by Musolff (Reference Musolff2016, Reference Musolff2021) has traced the deep historical roots of the NATION AS A BODY/PERSON metaphor, demonstrating its enduring influence on contemporary political thought. This framework maps the attributes of a living organism onto the state, facilitating discourse surrounding the nation’s health, maturity, strength and survival (Eriksen, Reference Eriksen, Hastrup and Olwig1997). These cognitive models serve as underlying conceptual scaffolds that guide reasoning (Zibin et al., Reference Zibin, Altakhaineh and Hamdan2022). Hence, by conceptualising the nation as a PERSON, political actors can evoke empathy, frame challenges as illnesses or injuries requiring treatment and present their policies as necessary interventions for the well-being of the national body. This rhetorical strategy has been observed across diverse contexts, from Chinese social media, where personification is employed to foster collective identity (Huang & Wang, Reference Huang and Wang2025), to European political history.
Equally foundational is the use of the NATION AS A FAMILY metaphor. Drawing on the seminal work of Lakoff (Reference Lakoff2002), scholars have shown how this frame structures the nation through the lens of kinship, carrying meaningful implications for hierarchy, belonging and mutual obligation (Lauenstein et al., Reference Lauenstein, Murer, Boos and Reicher2015; Polzenhagen, Reference Polzenhagen, Schröder, de Oliveira and Tenuta2022). The metaphor can evoke strong sentiments of loyalty and shared descent, casting citizens as siblings and leaders as paternal or maternal figures. This is evident in the use of terms like ‘Motherland’ to sacralise national territory in Russian discourse (Riabov, Reference Riabov2020) or in the framing of Turkish national identity through familial roles (Baruh & Popescu, Reference Baruh and Popescu2008). However, as Schatzberg (Reference Schatzberg2001) notes in his work on Central Africa, this metaphor can also legitimise patrimonial and authoritarian governance by casting the relationship between the state and its citizens as analogous to that of a father and his children.
Subsequent conceptualisations frame the nation in spatial and structural terms. The NATION AS A TERRITORY or container (Fox, Reference Fox2017) emphasises geographical boundaries as the basis for belonging, a frame that becomes particularly salient in discourses surrounding migration and national security (Pruitt, Reference Pruitt2019). The NATION AS A BUILDING or HOUSE metaphor meanwhile highlights themes of construction, stability and shared values (Feinberg & Wehling, Reference Feinberg and Wehling2018). This frame presents nation-building as a project requiring a solid foundation and collective effort, a common trope in post-conflict or transitional political rhetoric.
Although these organic and structural metaphors exhibit near-universal occurrence, research also highlights the significance of culturally specific frames that reflect unique historical and societal patterns (Polzenhagen, Reference Polzenhagen, Schröder, de Oliveira and Tenuta2022). Well-known examples include the MELTING POT and CULTURAL MOSAIC metaphors in the United States and Canada, which offer competing models for multicultural integration (Glazer & Moynihan, Reference Glazer and Moynihan1963; Siddiqui et al., Reference Siddiqui, Cimpian and Rutherford2020). Similarly, South Africa’s post-apartheid identity was famously encapsulated in the RAINBOW NATION metaphor (Bornman, Reference Bornman, Moeschberger and DeZalia2014), while Vietnamese diplomacy has been framed through the resilience of BAMBOO (Trinh & Vu, Reference Trinh and Vu2024). These metaphors are not interchangeable; they are embedded in local values and historical narratives, confirming that the conceptualisation of nationhood is inseparably bound to its cultural context.
Notwithstanding this rich body of scholarship, two limitations remain particularly relevant to the present study. First, research on NATIONHOOD metaphors remains concentrated largely on Western and East Asian contexts, leaving African multilingual and multi-ethnic settings comparatively underexplored. Second, many existing studies rely on smaller datasets or primarily qualitative interpretation. Taken together, these limitations suggest that while NATIONHOOD metaphors are widespread, their context-specific variation, particularly within African political discourse, remains insufficiently understood. The present study responds to this gap through a corpus-based, multifactorial analysis of NATIONHOOD metaphors in Nigerian presidential discourse. It therefore asks:
2.3. Research questions (RQs)
RQ1: What metaphors of NATIONHOOD are prevalent in Nigerian political discourse?
RQ2: What is the association between the extralinguistic factors of ethnicity, region and religion, and the selection of metaphor source domains for NATIONHOOD in Nigerian presidential discourse?
RQ3: How do these metaphor source domains and the extralinguistic factors of religion, ethnicity and region interact with one another?
3. Methodology
To address these questions, this study employs a corpus-based, quantitative methodology, integrating systematic metaphor identification with multifactorial statistical analysis.
3.1. Corpus description
The primary dataset for this study is the CORNIPS, a specialised, diachronic corpus compiled specifically for this research. CORNIPS comprises 358 speeches delivered by 16 Nigerian heads of state and presidents, spanning the 65-year period from Nigeria’s independence in 1960 to 2024. The corpus totals 875,481 words. Speeches were sourced from the National Archives of Nigeria, the National Library, digital newspaper archives and official government websites. Table 1 presents the composition of the corpus.
Corpus of Nigerian Presidential Speeches (CORNIPS) (1960–2024)
Table 1. Long description
The table contains six columns: S forward slash N, Names of military leaders, Positions, Periods, No. of speeches, and Word count.
* 1. Tafawa Balewa, Prime Minister, 1/10/1960 to 15/01/1966, 13 speeches, 19,204 words.
* 2. Gen Aguiyi-Ironsi, Military Head of State, 16/1/1966 to 29/7/1966, 9 speeches, 12,247 words.
* 3. Gen Yakubu Gowon, Military Head of State, 1/8/1966 to 29/7/75, 23 speeches, 51,335 words.
* 4. Maj Murtala Muhammed, Military Head of State, 29/7/1975 to 13/2/76, 7 speeches, 10,089 words.
* 5. Lt Gen Olusegun Obasanjo, Military Head of State, 13/2/1976 to 30/9/79, 23 speeches, 50,564 words.
* 6. Shehu Shagari, President, 1/10/1979 to 31/12/1983, 11 speeches, 30,626 words.
* 7. Gen Muhammadu Buhari, Military Head of State, 31/12/1983 to 27/8/85, 35 speeches, 51,992 words.
* 8. Brig Ibrahim B. Babangida, Military Head of State, 27/8/1985 to 26/8/93, 55 speeches, 189,339 words.
* 9. Chief Ernest Shonekan, Interim President, 26/8/1993 to 17/11/1993, 3 speeches, 14,922 words.
* 10. Gen Sani Abacha, Military Head of State, 17/11/1993 to 8/6/98, 12 speeches, 49,699 words.
* 11. Gen Abdulsalami Abubakar, Military Head of State, 8/6/1998 to 28/5/99, 4 speeches, 15,800 words.
* 12. Olusegun Obasanjo, President, 29/5/1999 to 29/5/2007, 43 speeches, 136,464 words.
* 13. Umaru Musa Yar’Adua, President, 29/5/2007 to 5/5/2010, 6 speeches, 10,295 words.
* 14. Goodluck Jonathan, President, 6/5/2010 to 29/5/2015, 40 speeches, 99,249 words.
* 15. Muhammadu Buhari, President, 29/5/2015 to 29/5/2023, 56 speeches, 106,988 words.
* 16. Bola Ahmed Tinubu, President, 29/5/2023 to 2024, 18 speeches, 26,668 words.
The total count for the corpus is 358 speeches and 875,481 words.
The corpus is compiled in Standard Nigerian English, a recognised variety of English, reflecting the status of English as Nigeria’s official language (Aboh & Chigbu, Reference Aboh and Chigbu2026). The communicative modality of these speeches is characterised by a blend of written-to-be-spoken and spoken formats (broadcasts), capturing the dynamic interplay between scripted rhetoric and oratory that typifies political communication in Nigeria.
3.2. Metaphor identification and source domain verification
The metaphorical analysis is theoretically grounded in Conceptual Metaphor Theory (Lakoff & Johnson, Reference Lakoff and Johnson1980). Lakoff and Johnson (Reference Lakoff and Johnson1980) defined conceptual metaphor as a correspondence between a target domain and a source domain, where the latter provides a more concrete conceptual framework for understanding the former. This study adopts a target-domain-driven approach, following Chigbu et al. (Reference Chigbu, Ukwunna and Aboh2026). This methodological framework necessitates a pre-selected target domain, thereby directing the analytical focus towards source-domain variation. Consequently, the terms ‘metaphor’ and ‘source domain’ are used interchangeably in this study.
The target domain of NATIONHOOD was identified through a set of keywords: nation (and its lemma nationhood), country, state, entity and land. A concordance search was performed for these keywords, and the surrounding linguistic context for each instance was manually analysed. The context window for the concordance search was set to the paragraph level to ensure that extended metaphorical mappings and multiword metaphors (e.g. ‘founding fathers’, ‘fabric of the nation’) were accurately captured in their discursive environment. Metaphorical expressions were identified using the Metaphor Identification Procedure Vrije Universiteit (MIPVU), a systematic and replicable protocol (Steen et al., Reference Steen, Dorst, Herrmann, Kaal, Krennmayr and Pasma2010). According to MIPVU, a lexical unit is identified as metaphorical if its contextual meaning is distinct from its more basic, concrete and human-oriented contemporary meaning.
Upon identification, each metaphorical keyword was categorised into a conceptual source domain (e.g. NATION AS A PERSON, NATION AS A JOURNEY). To ensure objectivity and replicability in this classification, a multi-step, corpus-assisted Source Domain Verification Procedure was employed (Ahrens & Jiang, Reference Ahrens and Jiang2020). This procedure triangulates evidence from large-scale linguistic resources, including the Suggested Upper Merged Ontology (SUMO), WordNet and online dictionaries, to verify the conceptual category of each keyword, thereby minimising researcher intuition. To address the challenge of overlapping boundaries between domains, strict criteria were established (which are available on the Open Science Framework). For instance, the PERSON domain was strictly reserved for biological, anatomical or life-cycle references (e.g. ‘sick’, ‘heal’, ‘mature’), the FAMILY domain was restricted to kinship terminology (e.g. ‘brother’, ‘fatherland’) and the PLANT domain was used for botanical growth descriptors (e.g. ‘nurture’, ‘fruit’). These delineations have been contextualised in previous metaphor studies (e.g. Chigbu & Ahrens, Reference Chigbu and Ahrens2025; Chigbu et al., Reference Chigbu, Ukwunna and Aboh2026).
3.3. Inter-rater reliability
To ascertain the reliability of the source domain coding, an inter-rater reliability test was conducted. Two trained analysts independently coded a random 10% subset of the data. While a 10% subset is standard practice, we acknowledge that given the 2,707 total tokens, a larger subset could provide even greater robustness. Cohen’s kappa scores indicated substantial to almost perfect agreement for the five dominant source domains included in this analysis: PERSON (k = .821), JOURNEY (k = .871), BUILDING (k = .871), FAMILY (k = .871) and WAR (k = .737), confirming the reliability of the coding scheme. The slightly lower kappa for WAR reflects the occasional difficulty in distinguishing literal military operations from metaphorical conflict in a country that experienced a civil war; discrepancies were resolved through discussion.
3.4. Variables description and log-linear analysis
To examine the relationship between leaders’ social identities and their metaphor choices, a four-way hierarchical log-linear analysis was conducted. This statistical method is ideal for analysing the relationships among multiple categorical variables, as it does not assume a dependent-independent variable structure and can model all main effects and higher-order interactions simultaneously (Gilbert, Reference Gilbert2022). The analysis was performed using IBM SPSS Statistics (v.31) on a dataset of 2,123 metaphorical tokens.
The four categorical variables integrated into the model were:
-
1. Source Domain (SD): This variable represents the metaphorical frame used. Based on a broader analysis of the corpus, the five most dominant and conceptually distinct source domains for NATIONHOOD were selected. The levels are: PERSON, JOURNEY, BUILDING, FAMILY and WAR.
-
2. Ethnicity (ETH): This variable represents the leader’s ethnic group. The six levels are: Hausa-Fulani, Yoruba, Igbo, Ijaw, Kanuri and Angas. The Hausa-Fulani category is a conventional grouping in Nigerian political analysis, reflecting the historical and cultural fusion of these two dominant northern groups.
-
3. Region (REG): This variable represents the leader’s geographical origin, a primary axis of political identity in Nigeria. The levels are: Northern and Southern.
-
4. Religion (REL): This variable represents the leader’s religious affiliation. The levels are: Christian and Muslim.
Table 2 provides a comprehensive summary of the leaders’ socio-demographic profiles.
Nigerian political leaders and social demographic information
Table 2. Long description
The table contains seven columns: S forward slash N, Names of military leaders, Positions, Religion, Ethnic identity, Region, and Gender. All 16 entries are listed as Male.
* 1. Tafawa Balewa: Prime Minister, Muslim, Hausa-Fulani, Northern.
* 2. Gen Johnson Aguiyi-Ironsi: Military Head of State, Christian, Igbo, Southern.
* 3. Gen Yakubu Gowon: Military Head of State, Christian, Angas, Northern.
* 4. Maj Murtala Muhammed: Military Head of State, Muslim, Hausa-Fulani, Northern.
* 5. Lt Gen Olusegun Obasanjo: Military Head of State, Christian, Yoruba, Southern.
* 6. Shehu Shagari: President, Muslim, Hausa-Fulani, Northern.
* 7. Gen Muhammadu Buhari: Military Head of State, Muslim, Hausa-Fulani, Northern.
* 8. Brig Ibrahim B dot Babangida: Military Head of State, Muslim, Hausa-Fulani, Northern.
* 9. Chief Ernest Shonekan: Interim President, Christian, Yoruba, Southern.
* 10. Gen Sani Abacha: Military Head of State, Muslim, Kanuri, Northern.
* 11. Gen Abdulsalami Abubakar: Military Head of State, Muslim, Hausa-Fulani, Northern.
* 12. Olusegun Obasanjo: President, Christian, Yoruba, Southern.
* 13. Umaru Musa Yar’Adua: President, Muslim, Hausa-Fulani, Northern.
* 14. Goodluck Jonathan: President, Christian, Ijaw forward slash Ogbia, Southern.
* 15. Muhammadu Buhari: President, Muslim, Hausa-Fulani, Northern.
* 16. Bola Ahmed Tinubu: President, Muslim, Yoruba, Southern.
4. Findings and discussion
The findings of the study are presented and discussed here in line with the research questions.
4.1. Metaphors of NATIONHOOD in Nigerian political discourse
This section addresses Research Question 1. The metaphor analysis reveals a total of 2,707 metaphorical keywords in CORNIPS, distributed across 27 source domains (SDs). Next, we examine metaphor dominance. Metaphorical dominance is operationalised as the extent to which specific source domains significantly shape the framing of a target concept, in this case, NATIONHOOD, based on their cumulative contribution to overall metaphorical productivity within a corpus. To determine the dominance of these SDs, we followed the criteria established by Ahrens and Zeng (Reference Ahrens, Zeng, Peters and Burridge2021), which define dominant metaphors as those accounting for up to 95% of the metaphorical tokens in a corpus, calculated based on cumulative percentages derived from total frequencies. While the 95% threshold is somewhat arbitrary, sensitivity tests indicated that lowering the threshold to 90% or raising it to 98% did not significantly alter the inclusion of the top five domains central to this analysis. Although the 95% threshold identifies a broader set of dominant domains, the subsequent log-linear analysis focuses on the five most frequent and conceptually distinct domains: PERSON, JOURNEY, FAMILY, BUILDING and WAR. Table 3 presents the list of the source domains, sample keywords and their frequencies.
Metaphor source domains of NATIONHOOD in CORNIPS
Table 3. Long description
The table consists of 14 ranked rows and a total row across six columns.
* Rank 1: P E R S O N. Keywords: beloved, life, survive, destiny, virile. Token frequency: 651. Cumulative frequency: 651. Cumulative percentage: 24.05.
* Rank 2: J O U R N E Y. Keywords: progress, move, forward, path, journey. Token frequency: 538. Cumulative frequency: 1,189. Cumulative percentage: 43.92.
* Rank 3: F A M I L Y. Keywords: fatherland, brothers, founding fathers, children, sons. Token frequency: 382. Cumulative frequency: 1,571. Cumulative percentage: 58.03.
* Rank 4: B U I L D I N G. Keywords: build, stable, strong, foundation, collapse. Token frequency: 372. Cumulative frequency: 1,943. Cumulative percentage: 71.78.
* Rank 5: W A R. Keywords: peace, section, confront, defend, save. Token frequency: 180. Cumulative frequency: 2,123. Cumulative percentage: 78.43.
* Rank 6: P L A N T. Keywords: develop, grow, thrive, viable, fruits. Token frequency: 135. Cumulative frequency: 2,258. Cumulative percentage: 83.41.
* Rank 7: B U S I N E S S. Keywords: interest, benefit, gains, enterprise, stake. Token frequency: 89. Cumulative frequency: 2,347. Cumulative percentage: 86.70.
* Rank 8: M A S T E R. Keywords: serve, work, duty, obligation, render. Token frequency: 68. Cumulative frequency: 2,415. Cumulative percentage: 89.21.
* Rank 9: P H Y S I C A L O B J E C T. Keywords: plunge, engulfed, fabric, indissoluble, lift. Token frequency: 59. Cumulative frequency: 2,474. Cumulative percentage: 91.39.
* Rank 10: S I C K N E S S. Keywords: plague, wounds, heal, sanity, afflict. Token frequency: 48. Cumulative frequency: 2,522. Cumulative percentage: 93.17.
* Rank 11: T E M P O R A L. Keywords: future. Token frequency: 47. Cumulative frequency: 2,569. Cumulative percentage: 94.90.
* Rank 12: M E T A L W O R K. Keywords: forging, weld, shape, recreate, reshape. Token frequency: 23. Cumulative frequency: 2,592. Cumulative percentage: 95.75.
* Rank 13: I M A G E. Keywords: image. Token frequency: 21. Cumulative frequency: 2,613. Cumulative percentage: 96.53.
* Rank 14: Others (15). Token frequency: 94. Cumulative frequency: 2,707. Cumulative percentage: 100.
* Total: Token frequency: 2,707.
The identified metaphor SDs in CORNIPS indicate that Nigerian political leaders rely heavily on figurative language to conceptualise Nigerian NATIONHOOD. Several of the employed metaphorical SDs align with those documented in extant literature, such as PERSON/BODY (e.g. Huang & Wang, Reference Huang and Wang2025; Musolff, Reference Musolff2021, Reference Musolff, Wuppuluri and Grayling2022) and FAMILY (e.g. Baruh & Popescu, Reference Baruh and Popescu2008; Lakoff, Reference Lakoff2002; Musolff, Reference Musolff2016). Additionally, some of these metaphor SDs are commonly used to conceptualise various political and social issues. In the next section, we examine the association between NATIONHOOD metaphors, particularly the top five metaphor SDs, and extralinguistic factors.
4.2. Associations between NATIONHOOD metaphor use and extralinguistic factors
This section addresses Research Questions 2 and 3 by examining whether leaders’ choices among the five most frequent source domains, PERSON, JOURNEY, FAMILY, BUILDING and WAR, are significantly associated with region, ethnicity and religion. The analysis focuses on 2,123 metaphorical instances and evaluates whether these variables correspond to systematic patterns of metaphorical choice.
4.2.1. Interaction of metaphor source domains, religion, ethnicity and region
To determine how the four variables interact, a hierarchical log-linear analysis with backward elimination was employed. This statistical procedure begins with a full (saturated) model containing all possible interactions. It systematically removes the least significant ones until only those that are statistically necessary to explain the observed data remain. This process reveals the underlying structure of the relationships between the variables. The model-fitting procedure progressed through eight steps to identify the most parsimonious model (see Field, Reference Field2024; Gilbert, Reference Gilbert2022; Tay, Reference Tay2016, for detailed introductions to log-linear analysis). The summary of this procedure is detailed in Table 4. The full SPSS output procedure can be accessed via OSF at https://osf.io/56sze/.
Step summary of backward elimination
Table 4. Long description
A table with six columns: Step, Generating class, Deleted effect, Chi-squared (chi super 2), d f (degrees of freedom), and Sig. (significance).
* Step 0: Generating class and Deleted effect are S D asterisk R E L asterisk asterisk R E G asterisk E T H. Chi-squared is .000, d f is 20, Sig. is 1.000.
* Step 1: Generating class is All 3-way interactions. Deleted effect is S D asterisk R E L asterisk asterisk R E G. Chi-squared is .000, d f is 4, Sig. is 1.000.
* Step 2: Generating class includes S D asterisk R E L asterisk E T H, R E L asterisk R E G asterisk E T H, and S D asterisk R E G asterisk E T H. Deleted effect is S D asterisk R E L asterisk E T H. Chi-squared is .026, d f is 20, Sig. is 1.000.
* Step 3: Generating class includes R E L asterisk R E G asterisk E T H, S D asterisk R E G asterisk E T H, and S D asterisk R E L. Deleted effect is S D asterisk R E G asterisk E T H. Chi-squared is .028, d f is 20, Sig. is 1.000.
* Step 4: Generating class includes R E L asterisk R E G asterisk E T H, S D asterisk R E L, S D asterisk R E G, and S D asterisk E T H. Deleted effect is S D asterisk R E G. Chi-squared is .000, d f is 4, Sig. is 1.000.
* Step 5: Generating class includes R E L asterisk R E G asterisk E T H, R E L asterisk S D, and S D asterisk E T H. Deleted effect is R E L asterisk R E G asterisk E T H. Chi-squared is .000, d f is 5, Sig. is 1.000.
* Step 6: Generating class is All 2-way interactions. Deleted effect is R E L asterisk R E G I O N. Chi-squared is .002, d f is 1, Sig. is .964.
* Step 7: Generating class includes R E L asterisk S D, S D asterisk E T H, R E L asterisk E T H, and R E G asterisk E T H. Deleted effect is R E L asterisk S D. Chi-squared is 6.897, d f is 4, Sig. is .141.
* Step 8: Generating class includes S D asterisk E T H, R E G asterisk E T H, and R E L asterisk E T H. Chi-squared is 6.926, d f is 78, Sig. is 1.000.
Abbreviations used: S D equals source domain, R E L equals religion, R E G equals region, and E T H equals ethnicity.
Note: Final model goodness of fit: likelihood ratio: (χ 2 (78) = 6.926, p = 1.000).
Abbreviations: SD = source domain, REL = religion, REG = region, ETH = ethnicity.
Table 4 illustrates the analytical steps towards the final model. The process began at Step 0 by testing the highest-order (four-way: SD*REL*REG*ETH) interaction. Its p-value of 1.000 indicated it was not significant and could be removed without loss of explanatory power. This iterative process continued, removing all three-way interactions (Steps 1–3) and several two-way interactions (Steps 4–7) because their removal, as indicated by their high p-values (all >.05), did not significantly worsen the model’s fit to the data. The procedure halted at Step 8 because deleting any of the remaining effects would have significantly worsened model fit. This process ensures that the final model is both accurate and parsimonious, that is, the simplest model that adequately explains the relationships in the data (Gilbert, Reference Gilbert2022). The goodness-of-fit for this final model is presented in the last row in Table 4.
The likelihood ratio chi-square value (χ 2 (78) = 6.926, p = 1.000) is non-significant. In log-linear analysis, a non-significant p-value (p > .05) indicates that the model is a good fit for the data, meaning there is no significant difference between the frequencies predicted by the model and the frequencies actually observed in the dataset (Field, Reference Field2024; Gilbert, Reference Gilbert2022). This confirms that the final model is an adequate representation of the relationships among the variables.
The backward elimination process determined that the most parsimonious model that fits the data consists of three significant two-way interactions. The generating class for this final model is:
-
1. [Source_Domain*Ethnicity]
-
2. [Region*Ethnicity]
-
3. [Religion*Ethnicity]
Of particular note, the analysis revealed that no three-way or four-way interactions were statistically significant. For instance, the potential three-way interaction of [Region*Ethnicity*Source_Domain] was eliminated during the model-fitting process. This finding indicates that the identified two-way relationships are stable and not contingent on the values of a third variable.
Moreover, while the study focuses on the influence of religion, region and/or ethnicity on metaphor use (source domain), the Region–Ethnicity and Religion–Ethnicity associations are treated as secondary associations that contextualise the primary Source Domain–Ethnicity relationship. This approach enables the study to explore how these factors interact to shape metaphorical preferences in a structured and contextual manner. The levels of these effects were tested using chi-square analysis, following Gilbert’s (Reference Gilbert2022) framework. In the following sections, we discuss each of these significant (primary) associations and evaluate the implications of the non-significant ones.
4.3. The primary association: Metaphor source domain and ethnicity
The optimal model identifies ethnicity as the primary variable, since it interacts significantly with source domain, region and religion. Most importantly, the significant [Source_Domain*Ethnicity] interaction indicates that leaders’ ethnic backgrounds are systematically associated with their metaphorical choices. Figure 1 illustrates the frequency distribution of the source domains across ethnic groups.
Frequency distribution of the source domains across ethnic groups.
Figure 1. Long description
The Y-axis is labeled Count and ranges from 0 to 300 with increments of 100. The X-axis is labeled Ethnicity and lists six groups from left to right: ANGAS, HAUSAFULANI, YORUBA, IGBO, IJAW, and KANURI. A legend titled SourceDomains identifies five categories: PERSON (light blue), JOURNEY (maroon), BUILDING (teal), FAMILY (dark purple), and WAR (dark green).
Data distribution by ethnic group:
* ANGAS: PERSON is highest at approximately 60. JOURNEY, BUILDING, and FAMILY are similar at around 45. WAR is lowest at approximately 20.
* HAUSAFULANI: Shows the highest overall counts. PERSON peaks at nearly 280, followed by JOURNEY at 210, BUILDING at 130, FAMILY at 125, and WAR at 85.
* YORUBA: PERSON is approximately 180, JOURNEY is 130, BUILDING is 95, FAMILY is 85, and WAR is 35.
* IGBO: Shows the lowest overall counts, with PERSON at approximately 20 and all other categories below 10.
* IJAW: FAMILY is the highest at 100, followed by JOURNEY at 95, PERSON at 70, BUILDING at 65, and WAR at 20.
* KANURI: JOURNEY is highest at 60, followed by PERSON at 45, BUILDING at 40, WAR at 25, and FAMILY at 20.
Nigerian leaders from different ethnic backgrounds exhibit distinct and predictable preferences for how they metaphorically frame the concept of NATIONHOOD. This relationship was found to be highly significant (χ 2 = 84.735, df = 20, p < .001). Given that this interaction [Source_Domain*Ethnicity] is central to our research question, we explore further to understand the nature of this interaction. Specifically, we examine the cross-tabulation of these two variables, paying close attention to the adjusted standardised residuals, which indicate where the observed counts deviate most significantly from what would be expected by chance. An adjusted standardised residual with an absolute critical value >1.96 is typically considered significant (Field, Reference Field2024; Tabachnick & Fidell, Reference Tabachnick and Fidell2013). Table 5 presents the cross-tabulation of source domain and ethnicity. The highlighted cells indicate a significant association.
Cross-tabulation of source domain and ethnicity
Table 5. Long description
A statistical table titled Cross-tabulation of source domain and ethnicity. The columns represent Source Domains: PERSON, JOURNEY, BUILDING, FAMILY, WAR, and a Total column. The rows represent Ethnicity: ANGAS, HAUSA-FULANI, YORUBA, IGBO, IJAW, KANURI, and a final Total row. Each ethnicity is subdivided into three metrics: Count, Exp (Expected count), and Adj. (Adjusted residuals).
Key data points include:
* ANGAS: Total count 213. Highest count in PERSON (62).
* HAUSA-FULANI: Total count 829. Highest count in PERSON (279).
* YORUBA: Total count 516. Highest count in PERSON (177).
* IGBO: Total count 31. Highest count in PERSON (18).
* IJAW: Total count 351. Highest count in FAMILY (99) and JOURNEY (96).
* KANURI: Total count 183. Highest count in JOURNEY (57).
The overall Total Count for all groups is 2123, distributed as: PERSON (651), JOURNEY (538), BUILDING (372), FAMILY (382), and WAR (180).
Statistics provided in the final column: chi-squared (20) = 84.735, p < .001, and Cramer’s V = 0.100.
Abbreviations: Exp = Expected count; Adj = Adjusted residuals.
4.3.1. Source Domain*Ethnicity
The adjusted residuals in Table 5 reveal several statistically notable rhetorical patterns associated with ethnicity. We discuss them in turn.
Ijaw leadership: The prominence of FAMILY and lower usage of PERSON
One notable pattern appears in the Ijaw category, represented in this dataset by President Goodluck Jonathan. The data reveal a preference for FAMILY metaphors (Adj. Residual = +5.5) and an underutilisation of PERSON metaphors (Adj. Residual = −4.8). This is particularly salient given Jonathan’s minority-group background and the political crises surrounding his presidency.
Example 1.
[a] My brothers and sisters, fellow citizens, we are all winners. In this context, there is no victor and no vanquished. (President Jonathan, 21 April 2011. IJAW ethnic group)
[b] Let me end this address at the point where I began. What matters most to all of us is Nigeria…irrespective of political parties or divide. What matters is Nigeria. This nation exists because we are one. We must, therefore, remain as one family …. (President Jonathan, 29 May 2012. IJAW ethnic group)
In Example 1a, delivered after his contentious 2011 victory, Jonathan’s opening ‘My brothers and sisters’ is a direct and strategic invocation of the FAMILY frame. It immediately attempts to deflate ethnic and political tensions by appealing to a higher-order, familial identity. The famous phrase ‘no victor and no vanquished’ functions here to signal the end of a fraternal dispute and the restoration of household harmony. In Example 1b, the strategy is even more explicit. By stating that Nigeria must ‘remain as one family’, he presents this conceptualisation as essential for national survival, using the metaphor as a tool to transcend partisan and ethnic loyalties.
The prevalence of the FAMILY metaphor in the discourse of the Ijaw leader provides strong evidence for a NATION-AS-KINSHIP cognitive model. This model foregrounds a conceptual structure based on shared descent, mutual obligation and internal harmony. Its core logic entails that national problems are ‘family disputes’ to be resolved through dialogue, and the leader’s role is that of a patriarch mediating between siblings.
In contrast, President Jonathan’s lower usage of the PERSON metaphor suggests a notable departure from a conceptualisation that frames the nation as a single, unitary body. For minority groups who have historically resisted political centralisation, the PERSON metaphor can carry negative connotations of forced assimilation and the erasure of distinct identities. By avoiding it, Jonathan subtly signals a respect for Nigeria’s diversity, reinforcing his image as a leader for all parts of the ‘family’. Crucially, the ‘Ijaw’ category in this dataset is represented solely by one individual. Therefore, while this preference for FAMILY metaphors reflects his specific rhetorical strategy, it cannot be broadly generalised to the entire Ijaw ethnic group without further empirical evidence.
Hausa-Fulani leadership: The prominence of PERSON and WAR metaphors
The Hausa-Fulani leaders exhibit a pattern that points to a security-oriented and disciplinary model of NATIONHOOD. They show a significant preference for both PERSON metaphors (Adj. Residual = +2.4) and WAR metaphors (Adj. Residual = +2.3), while strongly and significantly avoiding FAMILY metaphors (Adj. Residual = −2.8).
Example 2.
[a] Our intervention (coup) remains justifiable by the imperatives of securing the wholeness and oneness of our economy and nation. If we really understand the implications of the role of the military in our national agenda, then we can truly appreciate the full significance of the contribution of the military to our survival as a nation. (Gen Babangida, 22 January 1990. Hausa-Fulani)
[b] We are witnesses to the dangerous threat those frustrations posed to the very survival of our country as one nation. (Gen Abubakar, 1 October 1998. Hausa-Fulani)
Example 2 shows Generals Babangida and Abubakar conceptualising NATIONHOOD through PERSON-related keywords. In Example 2a, Babangida justifies military intervention by speaking of the nation’s ‘wholeness’, ‘oneness’ and ‘survival’, thereby framing Nigeria as a body whose integrity must be protected. The military’s role is consequently represented as an intervention required to preserve the national person-organism. In Example 2b, however, Abubakar’s appeal to the ‘survival’ of the country is located within a more transitional and retrospective discourse. His 1998 independence anniversary speech reviews Nigeria’s history of political crises, including the 1966 military intervention, the civil war, the collapse of the Second Republic and the failure of the 1993 elections to produce an elected government. In this context, the PERSON metaphor frames Nigeria as a vulnerable corporate body that has endured repeated political shocks and now requires stabilisation, reconciliation and institutional repair. And while both leaders rely on the PERSON metaphor, Babangida’s usage more strongly supports a command-and-control justification for military intervention, whereas Abubakar’s usage supports a caretaker image of restoring political order before the 29 May 1999 democratic handover. The co-occurrence of PERSON and WAR metaphors suggests a NATION-AS-ORGANISM-UNDER-THREAT scenario, in which the nation is framed as a corporate body whose integrity and survival require protection.
Yoruba leadership: Affective PERSON metaphors and attenuation of WAR
Yoruba leaders also show a significant preference for PERSON metaphors (Adj. Residual = +2.1), aligning them with their Hausa-Fulani counterparts in this specific choice. This suggests a shared tendency among the two largest ethnic groups to frame the NATION as a PERSON to be managed and cared for. However, Yoruba leaders employ significantly fewer WAR metaphors (Adj. Residual = −2.1) than expected by chance. From Example 3a–c, it is evident that Yoruba leaders’ prototypical metaphorical frame is often one of affective personification, where the Nigerian NATIONHOOD is framed as a beloved entity deserving of care.
Example 3.
[a] I consider this occasion very unique in the history of our dear country (Ernest Shonekan, Sep 1993. Yoruba)
[b] God is with us and together we shall make our dear country second to none in Africa and among the best in the world. (President Obasanjo, 29 May 2006. Yoruba)
[c] Our current circumstances should make us resolve to work better for the good of our beloved nation. (President Tinubu, 1 January 2024. Yoruba)
The recurring use of affectionate terms like ‘dear’ and ‘beloved’ in Example 3 conceptualises the Nigerian NATIONHOOD as a cherished person. This choice is particularly significant in Example 3a. Ernest Shonekan, a Yoruba man, was unexpectedly appointed interim president by a military general who had just annulled an election won by another Yoruba man, M.K.O. Abiola. In the moment of national crisis and ethnic grievance, framing Nigerian NATIONHOOD as a ‘dear country’ was a plea for unity, an appeal to a shared love for the nation that should transcend the political chaos.
The Yoruba pattern suggests that PERSON metaphors are not homogeneous: in this case, they often foreground affection and value rather than discipline, threat or emergency intervention.
Igbo leadership: PERSON metaphors and the unitary state
The analysis reveals a marked preference by Igbo leadership for PERSON metaphors (Adj. Residual = +3.3), primarily driven by the rhetoric of General Johnson Aguiyi-Ironsi. In Example 4a below, the phrase ‘well-being of the country’ directly invokes the PERSON metaphor by framing the Nigerian NATIONHOOD as a body that possesses health. In Example 4b, the nation is conceptualised as having a ‘destiny’, a singular life path, which is a core attribute of a person.
Example 4.
[a] It is not part of the policy of Government to curb the freedom of the Press but, at the same time, you will agree that it is the duty of the government to prevent any section of the community taking advantage of liberties which will not be conducive to the well-being of the country, its citizens and Government. (Gen Aguiyi-Ironsi, 21 February 1966)
[b] It will be necessary to review also the method of selection for future Ministerial appointments, in order that the destiny of the country will be placed in the hands of men capable of shouldering the heavy responsibilities of modern Government. (Gen Aguiyi-Ironsi, 21 February 1966).
Both examples support a unitary conceptualisation of the nation as a single entity with health, destiny and centralised direction. This is sharpened by the reduced usage of JOURNEY metaphors (Adj. Residual = −2.0). Within Ironsi’s short tenure, the goal was not to guide the nation on a journey but to fundamentally reconstitute it into one unified body.
Kanuri and Angas leadership: Contrasting Northern patterns
The Kanuri leader in the sample (Sani Abacha) shows a significant avoidance of FAMILY metaphors (Adj. Residual = −2.6), a pattern shared with his Hausa-Fulani counterparts. However, the Kanuri leader exhibits a strong tendency to prefer the JOURNEY metaphor (Adj. Residual = +1.9). In contrast, the Angas leader (Yakubu Gowon) stands out for having no statistically significant preferences or avoidances across the metaphor source domains.
From the foregoing, the findings suggest that the choice of a metaphorical source domain for Nigerian NATIONHOOD is not random but is systematically correlated with the speaker’s ethnic identity, likely reflecting the distinct historical experiences, cultural values and political objectives of these groups. In the following section, we explore the secondary associations generated by the model.
4.3.2. The secondary associations of social identity markers
The final model also confirmed two other significant two-way associations: [Region and Ethnicity] and [Religion and Ethnicity]. While these are not the primary focus of our study, their inclusion in the model is vital for two reasons. First, it validates the model’s robustness by confirming well-known demographic and sociopolitical realities of Nigeria. Second, it establishes ethnicity as the pivotal variable that links geography and religion to discourse patterns.
Region*Ethnicity
[Region*Ethnicity] (Partial Chi-Square = 2892.535, df = 5, p < .001): This association confirms that Nigeria’s major ethnic groups are predominantly concentrated in specific regions. The Yoruba are concentrated in the South-West, the Igbo in the South-East, and the Ijaw in the South–South, all together constituting Southern Nigeria, and the Hausa-Fulani, Angas, and Kanuri make up Northern Nigeria. This is a foundational feature of Nigerian society.
Religion*Ethnicity
[Religion*Ethnicity] (Partial Chi-Square = 2489.235, df = 5, p < .001): This equally strong association reflects the general religious composition of the ethnic groups in the dataset. The Hausa-Fulani and Kanuri leaders are predominantly Muslim, while the Igbo, Ijaw and Angas leaders are predominantly Christian. The Yoruba group is notable for having a significant number of both Christian and Muslim adherents, a fact also captured by the model.
The significance of these associations within the log-linear model indicates that, once a leader’s ethnicity is accounted for, their region and religion do not contribute additional explanatory power regarding their metaphor choices. This finding underscores the central conclusion: ethnicity serves as the primary determinant. This aligns with previous studies (e.g. Opone & Omosor, Reference Opone and Omosor2025; Tuki, Reference Tuki2024), which identified ethnic identification as a dominant variable that undermines any sense of national cohesion. This suggests that, in the present dataset, ethnicity is a more salient axis of political identification than religion or region.
Non-significant interactions
Equally significant to the findings of the model are the interactions it did not reveal. The systematic removal of higher-order interactions through the backward elimination process offers valuable insights by clarifying the absence of complex relationships.
No direct [Region*Source_Domain] or [Religion*Source_Domain] interaction
The analysis found no direct, significant association between a leader’s region (North vs. South) or religion (Christianity vs. Islam) and their choice of metaphor source domain. This is a crucial finding. It suggests that knowing only whether a leader is from the North or South, or whether they are Christian or Muslim, is not sufficient to indicate their metaphorical framing of NATIONHOOD. The lack of direct interaction also suggests that leaders from the same region, such as Southern Nigeria (Johnson Aguiyi-Ironsi, Olusegun Obasanjo, Goodluck Jonathan and Bola Tinubu), may not necessarily share the same metaphorical preferences or, by extension, similar political outlooks regarding the nation. The rhetorical differences observed in the data are not primarily structured along a simple North/South or Christian/Muslim divide. Instead, these factors are mediated through the more specific and cultural category of ethnicity.
No significant higher-order interactions
The absence of any significant three-way interaction, such as [Region*Ethnicity*Source_Domain], is particularly revealing. This means that the relationship we observed between ethnicity and metaphor choice is stable and consistent across different regions and religions. For example, the tendency for Hausa-Fulani leaders to favour PERSON and WAR metaphors does not significantly change based on other factors. Similarly, Jonathan’s preference for FAMILY metaphors is not moderated by region. This indicates that, within the model, the association between ethnic identity and metaphor choice is not substantially moderated by region or religion.
5. Discussion
The findings substantiate the premise that metaphor variation in Nigerian presidential rhetoric is associated with group-level social identity. This supports variationist approaches to CMT by showing that metaphorical preference is systematically associated with ethnicity rather than with region or religion alone (Kövecses, Reference Kövecses2012, Reference Kövecses2015). It moves beyond identifying universal metaphors to demonstrate how group-level social identity can correlate with cognitive-linguistic patterns in political discourse.
These results align with recent works on context-induced metaphor, which argue that metaphor production is shaped not only by embodied experience but also by situational and cultural context (Abdel-Raheem, Reference Abdel-Raheem2020; Kövecses, Reference Kövecses2015). The distinct rhetorical patterns we identified are linguistic manifestations of conceptual systems that have been entrenched through the shared historical, cultural and political experiences of each ethnic group. The stability of the [Ethnicity*Source_Domain] association suggests that leaders are drawing from a conventionalised, group-specific repertoire of conceptualisation, utilising distinct cognitive models (Zibin et al., Reference Zibin, Altakhaineh and Hamdan2022) to frame the abstract concept of NATIONHOOD.
The emergence of ethnicity as a key predictor of metaphorical choice should not be read as evidence of fixed ethnic cognitive styles. In this study, ethnicity functions as a proxy for shared sociopolitical history, rhetorical tradition and group positioning within the Nigerian state, rather than as a biological or deterministic category. The observed associations therefore indicate patterned rhetorical choices among political leaders from particular backgrounds, not uniform cognitive tendencies among all members of those groups. Furthermore, the non-significance of region and religion as direct predictors refines our theoretical understanding. It argues against using overly broad social categories in cognitive research and highlights the need for more micro-level, culturally sensitive models. Importantly, the data suggest that the lived experience of ‘ethnicity’ in Nigeria, a construct tied to language, kinship structures and specific historical grievances or advantages, offers a richer set of inputs for shaping cognitive models than the more diffuse categories of region or religion.
Furthermore, metaphorical selection is closely linked to the sociopolitical positioning of the speaker’s ethnic group. As Charteris-Black (Reference Charteris-Black2004) argues, the selection of a metaphorical source domain is never neutral but, in many cases, ideological. This is illustrated in the rhetorical strategies of the leaders in the corpus. President Jonathan’s reliance on the FAMILY metaphor can be read as a strategic frame to produce affective and political inclusion across ethnic and partisan divisions. Given Jonathan’s position as a minority-group leader from the Niger Delta, his use of the NATION-AS-KINSHIP model is an attempt to construct a pan-Nigerian consensus. This interpretation aligns with Semino et al.’s (Reference Semino, Demjén and Demmen2018) view that metaphors are not just cognitive reflections but active tools that shape social practice and political choice. In this context, Jonathan’s framing of the nation as a FAMILY appears to be a rhetorical strategy for moving beyond the ethnic arithmetic of Nigerian politics by appealing to a higher-order, fictive kinship capable of mitigating tensions during his presidency and securing him support for his re-election bid. This resonates with Sijadu’s (Reference Sijadu2020) account of African political discourse, where FAMILY metaphors operate as strategic manoeuvring by appealing to culturally shared assumptions about kinship, authority, obligation and collective belonging. The FAMILY metaphor therefore functions not only as a conceptual mapping but also as a rhetorical scenario in which political conflict is recast as intra-family disagreement requiring reconciliation rather than domination.
The Hausa-Fulani leaders have a preference for PERSON and WAR metaphors. These two metaphor source domains foreground a NATION-AS-ORGANISM-UNDER-THREAT model, in which the nation is imagined as a single biological entity whose continuity and integrity must be preserved. This supports the usefulness of moving beyond broad source-domain labels to the least schematic, since general mappings such as NATION AS BODY or NATION AS PERSON become politically meaningful through more specific evaluative and argumentative framings (Musolff, Reference Musolff2016, Reference Musolff2021, Reference Musolff, Wuppuluri and Grayling2022). The logic of this model allows the nation to be construed as ‘healthy’ or ‘sick’, or as vulnerable to hostile forces that threaten its integrity. Within this framing, the leader assumes the role of guardian, defender or physician, authorised to adopt strong, even painful measures to preserve the unity and stability of the national body. This interpretation is congruent with the military backgrounds of several leaders in this group, particularly its emphasis on discipline, protection and emergency intervention. The relative underuse of the FAMILY metaphor further supports this pattern, suggesting a limited reliance on relational or kinship-based models of governance and a stronger preference for hierarchical, protective and disciplinary conceptions of national leadership. However, this interpretation must remain cautious because many Hausa-Fulani leaders in the corpus were military rulers and their preference for PERSON and WAR metaphors may therefore reflect military background, regime type and historical moment as much as ethnic positioning.
By contrast, the Yoruba leaders’ preference for an affective PERSON metaphor, combined with their lower use of WAR metaphors, suggests a rhetorical emphasis on national well-being, institutional health and civic repair. This pattern contrasts with the Hausa-Fulani model, as it reflects a post-military political orientation more closely associated with democratic consolidation and socio-economic development than with militaristic struggle or emergency intervention, especially when read within the context of time and circumstance that brought most of the leaders from these two groups into power. The variation also demonstrates that PERSON metaphors do not constitute a single homogeneous category, since the same broad source domain can support different political meanings depending on its mapping principle (Ahrens, Reference Ahrens, Cameron, Deignan, Low and Todd2010) or scenario-based elaboration (Musolff, Reference Musolff2017). The logic underlying Yoruba leaders’ framing is therefore one of care, reconciliation and restoration: the nation is represented not primarily as an entity to be controlled or defended through force, but as a collective body to be loved, healed and strengthened.
However, the Igbo leadership’s preference for the unitary PERSON metaphor can be read in relation to General Johnson Aguiyi-Ironsi’s controversial Unification Decree, Decree No. 34 of 1966, which sought to replace Nigeria’s federal structure with a unitary system. In this context, conceptualising NATIONHOOD as a single, indivisible person becomes a rhetorical expression of a centralising political project. During Ironsi’s brief tenure, the dominant metaphorical orientation was not towards guiding the nation along a developmental journey, but towards reconstituting its political identity as one unified corporate body. However, because the Igbo category is represented by one leader and a short period of rule, this pattern should be interpreted as historically situated rather than as representative of Igbo political cognition more broadly. The dispreference for FAMILY metaphors by both Kanuri and Hausa-Fulani leaders suggests a common Northern leadership tendency to avoid familial frames. Furthermore, Sani Abacha’s (the Kanuri leader) preference for the JOURNEY metaphor may be interpreted as a rhetorical strategy to legitimise his rule, which began after the 1993 election crisis, by framing it as a necessary transition providing direction to move the nation forward out of the national crisis. In contrast, the neutral rhetorical profile of the Angas leader (Yakubu Gowon) aligns with his historical role as a reconciler who, after the civil war, sought to unify the country under the banner of ‘no victor, no vanquished’, perhaps carefully avoiding specific metaphorical frames that might alienate a section of the population.
5.1. Limitations
Although this study offers insights into correlations between extralinguistic factors and metaphorical choices, several methodological limitations merit consideration. First, the log-linear analysis reveals statistical associations, not causal relationships (Tabachnick & Fidell, Reference Tabachnick and Fidell2013). We can say that a leader’s ethnicity is strongly correlated with their metaphor choice, but we cannot definitively claim that ethnicity causes this choice. The relationship is likely complex and mediated by ideology, education, political strategy and individual psychology. As noted in the analysis, the preference for PERSON and WAR metaphors among Hausa-Fulani leaders may be confounded by the fact that many of these leaders were military generals operating within military regimes. Thus, temporal and cohort effects (military vs. civilian eras) present a strong alternative explanation for these patterns.
Secondly, the statistical model treats metaphor tokens as independent observations. This approach ignores the inherent clustering of data by individual speech and by speaker, which can potentially inflate the apparent statistical effects. Future research should reanalyse such data using mixed-effects (multilevel) models that include random intercepts for speaker and speech to ensure robust inferential validity. Additionally, the log-linear fit resulted in a p-value of 1.000, which, while indicating adequate model fit, may also point to sparse data or overfitting due to rare categories. Third, the dataset, while substantial, is skewed towards certain speakers and ethnic groups (e.g. Hausa-Fulani). The ‘Ijaw’ category, for instance, relies entirely on the speeches of a single president, meaning the findings reflect his personal rhetorical strategy rather than a broad ethnolinguistic generalisation. Finally, merging complex identities into broad labels like ‘Hausa-Fulani’ risks masking intra-group heterogeneity and shifting political alliances over the 65-year period.
6. Conclusion
This study investigated whether metaphorical conceptualisations of NATIONHOOD in Nigerian presidential rhetoric vary according to extralinguistic variables, particularly ethnicity, region and religion. Using log-linear modelling of CORNIPS, the analysis showed a significant association between metaphor source-domain choice and ethnicity, while region and religion did not independently predict metaphor choice. Ethnicity therefore emerged as the only macro-social variable in the model that directly interacted with metaphorical source-domain choice.
The key finding is that leaders from different ethnic backgrounds tend to draw on distinct cognitive models to structure and reason about the nation. For instance, the NATION-AS-FAMILY (KINSHIP) model used by minority leadership highlights a rhetorical strategy of inclusion. By contrast, the NATION-AS-PERSON/ORGANISM-UNDER-THREAT model favoured by historically dominant groups reflects a guardian/defender model, while the NATION-AS-A-CHERISHED-PERSON model illustrates a shift towards affective, non-combative framing. Collectively, these findings provide evidence that our conceptual systems are integrated with our social experiences and ethno-cultural histories.
Theoretically, this research contributes to our understanding of usage-based and variationist approaches to CMT. It shows that metaphor variation occurs not only at the level of broad source domains but also at the level of socially situated scenarios. Methodologically, it presents a framework for using corpus data to investigate the relationship between social categories and cognitive patterns. At the same time, because the study establishes a correlational rather than causal relationship, future research is needed to explore causality and account for data clustering through mixed-effects modelling. Experimental studies could test the persuasive impact of these different metaphorical frames on Nigerian citizens, while comparative work could apply this model to other pluralistic societies to determine whether ethnicity remains the dominant predictor of conceptual variation beyond the Nigerian context.
Data availability statement
The data and annotation manual that support the findings of this study are available on the Open Science Framework (OSF) at https://osf.io/56sze/.
Competing interests
The authors declare no competing interests.
Open access
