Semantic change in English loanwords in Polish: A diachronic study

Streszczenie

Niniejsza praca podejmuje problematykę zmian semantycznych zachodzących w angielskich zapożyczeniach funkcjonujących we współczesnej polszczyźnie. Celem badania jest diachroniczne prześledzenie ewolucji znaczeniowej szesnastu leksemów obcego pochodzenia w okresie od roku 1970 do 2020, z uwzględnieniem wpływu czynników pozajęzykowych na kierunek i tempo tych przemian. Za podstawę teoretyczną przyjęto klasyczne typologie zmian semantycznych — zwężenia, rozszerzenia oraz zmiany wartościujące — uzupełnione o koncepcje metaforyzacji i konotacyjnego przesunięcia znaczenia. W badaniu zastosowano metodę diachronicznej analizy korpusowej, obejmującej korpus liczący około 45 milionów tokenów, zorganizowany w pięć dekad, trzy rejestry i cztery domeny tematyczne, a ustalenia weryfikowano poprzez periodyzację słownikową. Wywód prowadzony jest od ogólnej charakterystyki procesów zapożyczania ku szczegółowej analizie poszczególnych grup leksykalnych. Stwierdzono, że w zapożyczeniach technicznych i profesjonalnych (komputer, broker, dżokej) dominuje zawężenie znaczenia, natomiast w leksemach związanych z kulturą i mediami (fan, lider, hit, imidż) obserwuje się jego rozszerzenie. Zmiany wartościujące (biznesmen, lobbysta, nerd) okazały się wrażliwe na przemiany polityczno-społeczne, zaś rozszerzenie metaforyczne wygenerowało innowacyjne użycia rodzime (surfować w znaczeniu nawigacji cyfrowej, profil jako tożsamość w sieci). Wszystkie trzy hipotezy badawcze zostały potwierdzone lub częściowo potwierdzone.

Słowa kluczowe: zapożyczenia angielskie w języku polskim, zmiana semantyczna, analiza diachroniczna, językoznawstwo korpusowe, metaforyzacja leksykalna

Abstract

This thesis investigates the semantic changes undergone by English loanwords that have become established in the Polish lexicon. The study traces the meaning evolution of sixteen borrowed lexemes diachronically across the period 1970–2020, examining how extralinguistic factors shape the direction and pace of semantic shift. The theoretical framework draws on classical typologies of semantic change — narrowing, broadening, and evaluative shift — supplemented by frameworks of metaphorical extension and connotative drift. The methodology combines diachronic corpus analysis of an approximately 45-million-token corpus spanning five decades, three registers, and four thematic domains, with dictionary-based periodisation to establish chronological benchmarks. The argument proceeds from a general account of borrowing processes to a close analysis of distinct lexical groups. Findings indicate that semantic narrowing predominates among technical and professional borrowings (komputer, broker, dżokej), while broadening characterises loanwords embedded in cultural and media discourse (fan, lider, hit, imidż). Evaluative changes affecting items such as biznesmen, lobbysta, and nerd prove sensitive to socio-political transformations, and metaphorical extension generates innovative native-Polish usages — most notably surfować reanalysed as digital navigation and profil reconceptualised as digital identity. All three research hypotheses are confirmed or partially confirmed by the corpus evidence.

Keywords: English loanwords in Polish, semantic change, diachronic analysis, corpus linguistics, lexical metaphorisation

List of Abbreviations

COHA

Corpus of Historical American English

KWIC

Key Word In Context

LWT

Loanword Typology

PELCRA

Poznań Corpus of Polish Texts

POS

Part-of-Speech

Introduction

The global diffusion of English in the post-war and post-Cold War periods has produced one of the most extensively documented episodes of lexical contact in the history of European languages. As English consolidated its position as the dominant medium of international commerce, technology, science, and popular culture, the vocabulary of recipient languages was subjected to sustained pressure from Anglophone sources. Polish, a West Slavic language characterised by rich inflectional morphology and a historically conservative orthographic tradition, has proved particularly susceptible to this pressure in the decades following the political and economic transformation of 1989. The rapid opening of Polish society to Western markets, media, and institutional models accelerated the influx of English loanwords into everyday usage, academic discourse, and professional registers at a rate that has attracted considerable scholarly interest. Estimates drawn from contemporary lexicographic surveys suggest that English currently constitutes the single most productive external source of new lexical items in Polish, surpassing the influence of Latin, French, and German that characterised earlier centuries of borrowing.

Scholarly engagement with English loanwords in Polish has followed several productive lines of inquiry. Phonological and morphological integration has received sustained attention, with researchers documenting the patterns through which English items are assimilated into the Polish declension system, the formation of hybrid derivatives, and the variable relationship between the degree of formal adaptation and the stability of an item within the lexicon ^[1]. Sociolinguistic investigations have examined the attitudinal dimensions of borrowing, exploring the prestige associations attached to Anglicisms in particular social groups and professional communities. Corpus-based frequency studies have mapped the distribution of English loanwords across genres and registers, contributing to a clearer picture of the quantitative scale of the phenomenon. What has remained comparatively underrepresented, however, is a systematic diachronic analysis of semantic change: the processes through which English loanwords, once established within Polish, develop meanings that diverge from, extend beyond, or narrow relative to their source-language equivalents. This gap is not merely of descriptive significance; it bears directly on theoretical questions concerning the mechanisms and directionality of contact-induced semantic change, which remain subjects of active debate within lexical typology and historical semantics ^[5].

The present study is designed to address that gap. Its primary objective is the systematic diachronic analysis of semantic changes undergone by a representative set of English loanwords in Polish across the period 1970–2020. Three operative hypotheses structure the investigation. The first posits that semantic narrowing constitutes the predominant change type in professional and technical registers, where borrowed items are recruited to denote specialised concepts and are consequently constrained in their referential range. The second proposes that metaphorical extension is concentrated in the domains of technology and popular culture, where the dynamics of rapid terminological expansion create conditions favourable to analogical meaning transfer. The third anticipates that ameliorative change tends to cluster in prestige-associated domains, while pejorative shift accompanies colloquial broadening into stigmatised social contexts. The analytical chapters test these hypotheses against corpus evidence and report the degree to which the observed data confirm, qualify, or complicate the predicted patterns.^{[29, s. 317]}

The empirical basis of the study is a purpose-built corpus of approximately forty-five million tokens, spanning five decades and structured across three registers — journalistic, academic, and colloquial written — and four thematic domains: technology, sport, economics, and lifestyle. The diachronic architecture of the corpus permits the reconstruction of meaning trajectories over time, enabling the analyst to establish when a given semantic shift first became attested, how rapidly it consolidated, and whether it remained confined to a particular register or migrated across the functional spectrum of the language. A systematic annotation scheme was developed for the identification and classification of semantic change types, and inter-annotator reliability was assessed using Cohen's kappa coefficient, which reached a value of 0.81, indicating substantial agreement and lending confidence to the consistency of the analytical framework across the full dataset .

The thesis is organised into three substantive chapters, preceded by this introduction and followed by a conclusion. Chapter One establishes the theoretical and historical foundations necessary for the analysis. It reviews the principal definitions and typologies of semantic change proposed within the relevant literature, covering processes of narrowing, broadening, amelioration, pejoration, and metaphorical extension as conceptualised by major theorists of lexical semantics ^[3]. The chapter further surveys the theoretical models of language contact and borrowing that inform the analytical framework, including Weinreich's classification of interference, Thomason and Kaufman's borrowing scale, and Haugen's model of loanword adaptation. A historical periodisation of English–Polish contact is also provided, distinguishing four broad phases: the pre-1900 period of limited trade and sporting vocabulary; the interwar and wartime decades in which military and technical terminology was absorbed; the post-1989 era of economic transformation; and the digital and social media period that extends into the present.

Chapter Two sets out the methodological apparatus in full. It details the construction and composition of the corpus, the procedures employed for the identification of loanwords and their semantic values at successive points in time, and the classification criteria applied to each attested change. The chapter addresses the principal methodological challenges encountered in diachronic semantic analysis of borrowed vocabulary, including the attestation problem that arises when early usage is sparsely documented, the difficulties involved in defining semantic identity across time, and the complications introduced by register reassignment and shifts in the English etymon itself. The operative hypotheses stated in the introduction are elaborated in Chapter Two with reference to the specific variables and conditions under which each is expected to hold, and the quantitative summarisation procedures used in Chapter Three are described and justified.

Chapter Three presents the core empirical findings. Sixteen English loanwords are subjected to detailed diachronic analysis, organised according to the dominant change type each exemplifies. Section 3.1 examines narrowing through the cases of komputer, dżokej, menedżer, and broker. Section 3.2 addresses broadening in the items fan, lider, hit, and imidż. Section 3.3 considers evaluative change — both ameliorative and pejorative — as observed in biznesmen, lobbysta, nerd, and establishment. Section 3.4 analyses metaphorical extension through klips, surfować, hosting, and profil. Section 3.5 synthesises the findings across chronological, domain-based, divergence, and typological dimensions, and provides a formal assessment of each of the three hypotheses in light of the evidence.

The concluding chapter draws together the principal findings of the study. It confirms the first and second hypotheses in full and offers a qualified assessment of the third, observing that evaluative change proved more sensitive to the political valence of specific historical moments than to domain membership alone — a finding that introduces a sociohistorical dimension absent from the original hypothesis formulation. The conclusion also identifies four directions for future research: the extension of the corpus to include spoken and digital registers; the application of computational methods for large-scale semantic tracking; a comparative study situating Polish within the broader Slavic context; and a longitudinal follow-up designed to capture developments in the post-2020 period.

The significance of the present study operates on two levels. At the theoretical level, it contributes evidence concerning the mechanisms and directionality of contact-induced semantic change to a field in which the empirical base, particularly for Slavic recipient languages, remains thinner than for Western European cases. The findings bear on longstanding debates regarding whether borrowing languages tend to narrow, stabilise, or expand the meanings of items they adopt, and whether the outcome varies predictably with register, domain, or historical period ^[5]. At the practical level, the study has implications for Polish lexicography, where the documentation of borrowed vocabulary frequently lags behind attested usage; for translation practice, where false cognates and drifted meanings between Polish and English Anglicisms constitute a recognised source of error; and for the teaching of Polish as a foreign language and of English to Polish speakers, both of which require accurate accounts of where the semantic values of shared items converge and diverge.

A number of delimitations govern the scope of the investigation and should be stated explicitly. The corpus is restricted to written sources; spoken language, digital communication, and informal online registers fall outside the study's scope, though their relevance to contemporary semantic change is acknowledged and proposed as a priority for subsequent research. The grammatical coverage of the analysis is confined to nouns and verbs, which together constitute the dominant word classes among English loanwords in Polish and offer the richest documentation across the decades under study; adjectival and adverbial borrowings are noted where relevant but are not subjected to systematic analysis. The temporal window of 1970–2020 was selected to ensure comparability of data volume across decades and to include the pivotal post-1989 acceleration of borrowing while maintaining a defensible historical baseline in the pre-transformation period. Finally, the four thematic domains — technology, sport, economics, and lifestyle — were chosen on the basis of their documented productivity as channels of Anglicism transfer into Polish and their representation across all three registers included in the corpus. These delimitations are intended not to diminish the study's claims but to define clearly the conditions under which those claims are warranted and to delineate the terrain that remains open for future investigation.

It is hoped that the analysis presented in the following chapters will demonstrate that English loanwords in Polish do not function as stable replicas of their source-language counterparts but instead acquire independent semantic lives shaped by the linguistic, cultural, and historical context of the recipient language. This process of semantic individualisation, documented here across five decades and four domains, constitutes a phenomenon of considerable interest both for the theory of lexical borrowing and for the broader understanding of how contact between languages transforms the conceptual resources available to their speakers.

Chapter 1. Theoretical and Historical Foundations of Loanword Semantics

1.1. Defining Semantic Change: Typologies, Mechanisms, and Terminological Debates

Semantic change — broadly understood as the evolution of word meanings over time — stands as one of the central concerns of historical and contact linguistics. The process through which a lexical item acquires, loses, narrows, or expands its range of reference has attracted sustained scholarly attention since the nineteenth century, and its systematic study has yielded a rich, if not always consistent, theoretical vocabulary. Foundational contributions to the field identified a core set of change types that continue to organise contemporary research: widening of meaning (generalisation or broadening), narrowing of meaning (specialisation or restriction), amelioration, pejoration, metaphorical extension, and metonymic shift ^{[8, p. 4145]}. These categories, while heuristically useful, are not mutually exclusive, and individual lexical items frequently exhibit overlapping or sequential change processes across documented historical periods ^{[9, p. 1380]}.

The causes underlying semantic change have been classified in multiple, partially complementary frameworks. Ullmann identified six principal causal factors: linguistic, historical, social, psychological, foreign influence, and the need for new nomenclature ^{[2, p. 116]}. Of these, the factor of need is considered especially operative in loanword semantics, where restriction and extension processes frequently respond to communicative gaps in the recipient language ^{[2, p. 116]}. Hasan similarly observes that words do not convey fixed, context-independent meanings; rather, they are invested with meaning according to the totality of the communicative context, which explains why identical forms may develop divergent semantic profiles across languages ^{[9, p. 1380]}. The interconnection between synchronic processes of meaning extension and diachronic processes of semantic change is thus not incidental but structural, since the same mechanisms that produce contextual variation at a given moment may crystallise, over time, into distinct historical stages of a word's semantic evolution ^{[9, p. 1381]}.

Empirical studies of semantic change in borrowed vocabulary have produced quantitative distributions that reveal important asymmetries in the prevalence of change types. An analysis of English loanwords in Indonesian media found that 73 per cent of examined items exhibited no change of meaning, while narrowing of meaning accounted for 17 per cent and widening for 10 per cent of the data; regeneration, degeneration, metaphor, and metonymy were each represented at lower frequencies ^{[1, p. 405]}. A comparable investigation of Malay words of Sanskrit origin shared with Thai found that, among fifty sampled items, semantic widening was the most prevalent type (24 instances), followed by semantic narrowing (22 instances) and meaning transfer (4 instances), suggesting that generalisation may be the historically dominant tendency in long-established contact situations ^{[8, p. 4149]}. These distributions are, of course, sensitive to the methodological choices governing corpus construction and the theoretical criteria used to distinguish one change type from another.

The study of semantic change in borrowed vocabulary is further complicated by persistent terminological ambiguity in the scholarly literature. A fundamental distinction must be drawn between borrowing as a process and loanword as its product: the former denotes the act of transfer, the latter the resulting lexical unit that has been adopted with little to no modification ^{[6, p. 43]}. Despite the apparent clarity of this distinction, the two concepts are widely employed as interchangeable terms in scholarly works, generating analytical confusion that is not merely terminological but substantive ^{[6, p. 43]}. Maslov and Kornieva argue that the absence of dedicated, universally accepted terms for the two aspects of the phenomenon — the process and its result — undermines cross-linguistic comparison and impedes the development of unified theoretical frameworks ^{[6, p. 45]}. The present study adheres to the terminological distinction between borrowing (process) and loanword (unit), and treats semantic change as a property of the unit that may or may not have accompanied, and may or may not continue after, the initial act of borrowing.

A further axis of theoretical debate concerns the distinction between internally driven and contact-induced semantic change. Internally driven change arises from mechanisms such as speaker innovation, analogical extension, taboo avoidance, frequency effects, and pragmatic inference, whereas contact-induced change is precipitated by interaction with another linguistic system. In loanword semantics, the two mechanisms are frequently co-operative rather than exclusive: the entry of a borrowed form may trigger semantic reorganisation in pre-existing lexical fields of the recipient language, while simultaneously the borrowed item itself undergoes semantic adjustment under pressure from the recipient system ^{[11, p. 183]}. Pyles observed, as cited in Al-Athwary, that while semantic change is frequently unpredictable, it is not wholly chaotic ^{[2, p. 116]} — a characterisation that captures the tension between the systematicity sought by theoretical models and the irreducibly contextual nature of individual semantic trajectories. It is this productive tension between regularity and contingency that motivates the present study's combination of theoretical classification with corpus-based diachronic analysis.

Table 1.1. Principal types of semantic change in borrowed vocabulary

Type	Alternative designations	Definition	Representative example
Widening (broadening)	Generalisation, extension	The meaning of a word becomes broader than in the source language	Sanskrit karma → Malay 'fate, destiny' [8, p. 4145]
Narrowing (restriction)	Specialisation	The meaning of a word becomes more restricted than in the source language	English loanwords in Korean restricted to technical domains [5, p. 32]
Metaphorical extension	Semantic transfer, shift	A word acquires new meaning via metaphorical mapping onto a related domain	English verbs borrowed into Korean with transferred meanings [5, p. 32]
Amelioration	Elevation	The evaluative component of meaning shifts towards positive connotation	Prestige-driven borrowings across multiple donor languages [7, p. 3]
Pejoration	Degradation, degeneration	The evaluative component shifts towards negative connotation	Degeneration type attested in Indonesian loanword corpus [1, p. 405]
No semantic change	Semantic stability	The meaning of the borrowed item remains equivalent to the source form	73% of English loanwords in Indonesian e-paper corpus [1, p. 404]

1.2. Language Contact and Borrowing: Theoretical Models

The theoretical modelling of lexical borrowing under conditions of language contact draws upon a tradition that is both rich and contested. Among the foundational contributions, Haugen's importation-substitution model, articulated in his seminal 1950 paper, remains a persistent point of reference. Haugen distinguished between loanwords proper — in which both the morphemic content and phonological substance of the source form are imported — and loan-blends as well as loan-translations, in which the meaning is borrowed but the phonological material is supplied by native resources ^{[3, p. 163]}. Haugen's insistence that phonological adaptation is not haphazard but follows systematic substitution of source phonemes by the nearest available segments in the recipient system established a principle that has guided subsequent work in loanword phonology and morphology alike ^{[3, p. 163]}. Two questions remain unresolved in his wake: what precisely makes one segment the „most similar” native equivalent of a foreign segment across typologically distinct systems, and whether the substitution process is more accurately characterised as a phonological or a phonetic operation ^{[3, p. 163]}.

A theoretically influential distinction that cuts across phonological and semantic dimensions is that between cultural borrowing and replacement borrowing. Cultural borrowings expand the vocabulary of the recipient language by designating concepts or objects that were previously absent from it, whereas replacement borrowings displace lexical items already present in the recipient system ^{[7, p. 3]}. The causality of replacement borrowing is more complex, as it requires explanation of why a native form is dislodged by a borrowed alternative; among the factors identified are cultural salience, functional utility, and prestige — the last of which is characterised as a relative and historically situated notion, operative at scales ranging from interpersonal interaction to the global diffusion of dominant languages ^{[7, p. 3]}. Carling and colleagues found preliminary evidence of a negative correlation between lexical replacement and the abstract borrowability of an item, suggesting that items most easily borrowed are not necessarily those most likely to displace native equivalents ^{[7, p. 11]}.

The study of borrowing profiles across languages has been substantially advanced by large-scale typological projects. The Loanword Typology (LWT) project, which encompasses data from forty-one languages across a deliberately global sample, constructed a meaning list designed to provide an empirically grounded basic vocabulary of high stability, enabling cross-linguistic comparison of borrowing tendencies across semantic domains ^{[7, p. 19]}. Research drawing on this and related resources has shown that loanword layers in the basic vocabulary of a language provide an adequate cross-section of the full borrowing profile, though prehistoric contacts tend to be more strongly represented in basic vocabulary than recent ones, since recent loanwords frequently occupy cultural rather than basic-vocabulary domains ^{[12, p. 55]}. The uneven distribution of loanwords across semantic fields — with technology, religion, and modern-world vocabulary exhibiting the highest borrowing rates — reflects the cultural and historical specificity of contact situations, as documented across numerous contact pairings ^{[10, p. 956]}.

The cognitive dimension of borrowing has been foregrounded in research on bi-dialectal and bilingual speakers as agents of lexical innovation. Wu and colleagues demonstrated that bi-dialectal speakers tend to create separate lexical representations for encountered dialectal variants and to draw on etymologically related morphemes when forming borrowed compounds, whereas monolingual or monolectal speakers are more likely to assimilate novel forms to existing patterns ^{[13, p. 1]}. This finding supports the view that the bilingual speaker serves as the primary locus of semantic negotiation during the initial phases of borrowing, since it is at the individual cognitive level that the mapping between source and recipient systems is first established ^{[3, p. 172]}. Where bilinguals are few in a community, non-phonological factors such as phonetic approximation, orthographic conventions, and analogical pressure play a correspondingly larger role in determining both the form and the meaning of the borrowed item ^{[3, p. 172]}.

The semantic dimension of borrowing has been approached through a variety of analytical frameworks, including semantic field theory, which holds that loanwords entering a recipient-language lexical field without corresponding native terms simply fill a vacancy, whereas loanwords that compete with existing terms are predicted to induce semantic reorganisation across the field ^{[5, p. 33]}. Chinese loanword scholarship has identified five principal borrowing strategies — phonetic translation (transliteration), semantic translation (loan translation), hybrid borrowing, calquing, and borrowings retaining Latin-letter elements — illustrating that the formal pathway of borrowing has significant implications for the semantic outcome, since purely phonetic adaptation tends to import meaning wholesale, while loan-translation and calquing involve semantic reanalysis in the source language ^{[14, p. 1909]}. The interplay between formal strategy and semantic outcome is a recurring theme across contact-linguistic research and is directly relevant to the analysis developed in subsequent chapters of the present study.

1.3. Phonological and Morphological Integration of English Loanwords in Polish

The formal integration of borrowed lexical items into the phonological and morphological systems of the recipient language is a well-documented but theoretically complex process. Haugen's founding insight — that phonological substitution tracks the distance between the phonological systems in contact and the degree of bilingual competence in the borrowing community — has proved durable across a wide range of contact situations ^{[3, p. 163]}. The framework of Feature Geometry addresses the question of segmental similarity by operationalising it in terms of shared feature nodes, turning Haugen's intuitive notion of „most similar” into a formally computable relationship; on this account, the „closest” native segment is the one that shares the greatest number of feature-geometric nodes with the source segment ^{[3, p. 172]}. This formalisation has substantially advanced the theoretical account of why loanword phonology is systematic rather than idiosyncratic, though it leaves open the question of how competing feature-geometric distances are resolved when a single source segment has multiple near-equivalents in the recipient system.

Cross-linguistic evidence reveals that positional factors may override source-segment identity in determining the form of a phonological adaptation. Korean loanword phonology provides a striking illustration: English /l/ and /r/ are both adapted in Korean, but which reflex appears depends on syllabic position rather than on the identity of the source phoneme — in onset position the reflex surfaces as [ɾ], while in coda position it surfaces as [l], in conformity with Korean distributional constraints ^{[3, p. 170]}. This positional sensitivity demonstrates that adaptation is governed by the structural grammar of the recipient language and not merely by phonetic similarity judgements. In Polish, the structural context is defined by a West Slavic typological profile that combines rich inflectional morphology, high tolerance for complex consonant clusters at syllable boundaries, and fixed penultimate stress — features that collectively determine the nativisation pathways available to incoming English loanwords. Items such as komputer, dżinsy, weekend, and marketing illustrate the range of adaptation outcomes, from near-transparent phonological correspondence to systematic substitution of non-native segments.

The degree of formal integration achieved by a loanword is not uniform across the lexicon but correlates with a range of factors including age, frequency of use, pragmatic context, and discourse register. Evidence from Imbabura Quechua shows that 88 per cent of Spanish loanwords were fully or almost fully assimilated to native phonological patterns, with only 7 per cent remaining unintegrated; integration was found to be conditioned by the interaction of age, frequency, pragmatics, and discourse ^{[10, p. 958]}. The Chinese loanword system provides further illustration of the relationship between formal integration and semantic productivity: transliterated elements have, over time, developed identifiable morphemic meanings, and certain borrowed letter sequences have become productive morphemes capable of generating new lexical items ^{[14, p. 1912]}. This morphemisation of phonetically adapted material represents an advanced stage of integration that certain English loanwords in Polish have begun to exhibit, particularly in technolectal and youth-language registers where derivational productivity is well documented.

Morphological integration into Polish inflectional paradigms presents its own analytical challenges. English nouns entering Polish must be assigned grammatical gender, integrated into one of the nominal declension patterns, and — in the case of verbs — fitted to an aspectual and conjugational paradigm. The outcome of this process is not always predictable from formal criteria alone: semantic factors, analogy with formally similar native items, and sociolinguistic prestige effects all play a role. Hybrid formations — structures that combine morphological material from both the source and recipient language — represent a particularly productive intermediate stage, attested across numerous contact languages including Russian, where English loanwords have generated hybrid morpheme-structure words whose lexical-semantic motivation is not transparent to speakers lacking knowledge of the source language ^{[11, p. 187]}. The relationship between degree of formal integration and semantic stability is theoretically significant for the present study: items that remain formally marked as foreign may retain a narrower, more technically bounded meaning, while thoroughly nativised items are more susceptible to semantic broadening under the influence of recipient-language associative and derivational patterns.

• Phonological adaptation involves substitution of source phonemes by the nearest available segment in the recipient system, governed in principle by Feature Geometry ^{[3, p. 172]}
• Positional constraints in the recipient language may override source-segment identity, as demonstrated by Korean onset-coda alternations for English /l/ and /r/ ^{[3, p. 170]}
• Morphological integration in Polish requires gender assignment, inflectional paradigm membership, and, for verbs, aspectual classification
• Hybrid formations combine morphological material from source and recipient languages, producing structures of mixed formal motivation ^{[11, p. 187]}
• Degree of phonological-morphological integration correlates with semantic stability and susceptibility to subsequent semantic change ^{[10, p. 958]}

1.4. A Historical Outline of English–Polish Language Contact

The history of English-Polish language contact is best understood as a sequence of distinct phases, each characterised by specific vectors of transmission, particular semantic domains of borrowing, and specific sociolinguistic attitudes towards the incoming lexical material. While the present chapter confines itself to a broad outline, the historical periodisation established here provides the necessary context for the diachronic analysis conducted in Chapter 3. It is worth noting at the outset that Polish contact with English — while increasingly significant — has historically been less intensive than contact with Latin, French, German, or Russian, each of which has contributed substantially larger strata to the Polish lexicon; English influence has accelerated dramatically in the post-1945 period and most emphatically following the political and economic transformations of 1989. The asymmetry is visible from both directions: studies of Polish loanwords in English identify only approximately twenty direct borrowings of Polish origin attested in the Oxford English Dictionary, confirming that the contact relationship has been fundamentally unidirectional ^{[4, p. 132]}.

The first period of contact, spanning the pre-twentieth-century era, was characterised by sporadic lexical exchange mediated by maritime trade, sporting culture, and the prestige associated with British civilisation. Items entering Polish from English during this period tended to reflect institutional and cultural realities specific to the British context: titles, forms of address, social categories, and sporting terminology. The mechanisms of transmission were largely indirect, with French and German frequently serving as intermediary languages through which English items reached Polish — a pathway that complicates etymological attribution and implies that the phonological and semantic form of early anglicisms was shaped by literate intermediation rather than by the productive bilingual norms that govern more direct contact situations. The relative infrequency of direct contact between English and Polish speakers during this period sets it apart structurally from the later phases of intense, immersive exposure.

The second period, encompassing the interwar years and the Second World War, was marked by intensified contact through military cooperation, particularly following the formation of Polish Armed Forces in the West and the substantial Polish community that settled in Britain after 1945. Military and technological terminology, together with the vocabulary of aviation and naval operations, constitutes the characteristic borrowing domain of this period. The post-war Polish diaspora in Britain and North America also served as a channel for more diffuse cultural and colloquial anglicisms that entered Polish through correspondence, returned emigrants, and eventually broadcast media. The semantic fields opened by this period of contact overlap significantly with those documented in studies of English influence on other Slavic languages: Alyunina and Nagel, investigating 487 English loanwords in Russian, found that technological, cultural, and institutional vocabulary constituted the largest categories, with three interaction types observable between the verbal code of the host language and the incoming lexical material ^{[11, p. 176]}.

The third and most consequential period was inaugurated by the political and economic transformations of 1989, which removed ideological barriers to the adoption of Western concepts and simultaneously exposed Polish society to unprecedented volumes of English-language cultural, commercial, and technological content. The influx of anglicisms associated with this transformation was characterised by the simultaneous entry of large numbers of items across multiple semantic domains — information technology, business, marketing, finance, youth culture, and popular media — within a compressed historical timeframe. A structurally parallel dynamic was observed in the Chinese context, where the Reform and Opening-up policy removed ideological constraints and precipitated a large-scale influx of transliterated loanwords for which appropriate semantic equivalents were initially unavailable ^{[14, p. 1910]}. In both contexts, the absence of unified normative standards governing loanword adaptation resulted in considerable inconsistency in borrowing practices and ongoing standardisation debates among linguists, lexicographers, and language planners ^{[14, p. 1912]}.

The fourth and current period, defined by digital communication and social media, has accelerated and diversified the channels through which English items enter Polish. The immediacy of online communication means that new anglicisms may achieve widespread use within months of their first appearance in English-language sources, compressing the timescale of borrowing and integration in ways that challenge methodological frameworks designed for the study of more gradual historical change. Sociolinguistically, the key agent of this phase is the speaker who commands substantial passive competence in English without full productive bilingualism — a group that corresponds structurally to the communities described in research on Quebec French loanwords, where the relative scarcity of highly proficient bilinguals led to greater influence of non-phonological factors such as orthography and analogy in shaping loanword forms ^{[3, p. 172]}. The stratification of attitudes towards anglicisms in Polish — ranging from enthusiastic adoption in technological and commercial discourse to purist resistance in academic and official registers — replicates a pattern observable across multiple recipient-language communities and constitutes a key explanatory variable for the semantic divergences that the subsequent chapters of the present thesis document and analyse.

Table 1.2. Periodisation of English–Polish language contact

Period	Approximate dates	Principal contact vectors	Dominant borrowing domains
Pre-twentieth century	Before 1900	Maritime trade, sport, literate intermediation	Institutional titles, sporting terminology
Interwar and wartime	1918–1945	Military cooperation, diaspora formation	Military, aviation, and naval vocabulary
Post-1989 transformation	1989–2000s	Market economy, broadcast media, technology	IT, business, marketing, youth culture
Digital and social media era	2000s–present	Online communication, global media platforms	Social media, gaming, e-commerce registers

• English has historically exerted less influence on Polish than Latin, French, German, or Russian; the reverse direction likewise shows a small imprint, with approximately twenty Polish-origin entries identified in the Oxford English Dictionary ^{[4, p. 132]}
• Pre-twentieth-century English borrowings in Polish were frequently mediated through French or German as intermediary languages, shaping their phonological and semantic form accordingly
• The post-1989 period produced large-scale simultaneous borrowing across multiple semantic domains, paralleling the dynamics observed in post-reform Chinese ^{[14, p. 1910]}
• The current digital era compresses borrowing timescales and elevates the role of partially bilingual speakers in shaping loanword forms ^{[3, p. 172]}
• Sociolinguistic attitudes to anglicisms in Polish are stratified by register, generation, and institutional context, creating systematic variation in adoption rates and semantic stabilisation trajectories

Chapter 2. Methodology, Corpus, and Analytical Framework

2.1. Research objectives, hypotheses, and scope

The present study is grounded in a clearly delineated epistemological framework, the articulation of which constitutes the initial task of this chapter. The primary research objective is defined as the systematic identification, documentation, and classification of semantic changes undergone by selected English loanwords in Polish across a chronological span extending from approximately 1970 to 2020. Particular attention is directed to three dimensions of each attested change: its direction (whether the shift constitutes narrowing, broadening, amelioration, pejoration, or metaphorical extension), its degree (whether the original sense is retained alongside the new one or displaced by it), and its register-specificity (whether the change is confined to a particular genre or distributed across several). The choice of a fifty-year window is motivated by the ambition to capture at least two full cycles of borrowing dynamics in Polish: the final decades of the socialist period, during which English loanwords entered the language primarily through technical and scientific channels, and the post-1989 era of accelerated contact, during which simultaneous adoption across multiple semantic domains became the norm. Secondary objectives include the establishment of statistically observable tendencies in the assembled corpus — such as the relative frequency of narrowing versus broadening, or the concentration of evaluative shifts within specific lexical domains — and the situating of these tendencies within existing theoretical models of semantic change and language contact as reviewed in Chapter 1.

The operative research hypotheses of the study are three in number. The first proposes that semantic narrowing constitutes the most prevalent mechanism of semantic change among English loanwords entering specialised professional registers in Polish, on the grounds that such registers impose domain-specific constraints that select for and stabilise restricted senses while peripheral meanings remain inaccessible to non-specialist users. The second hypothesis holds that metaphorical extension is disproportionately frequent among loanwords adopted in the domains of technology and popular culture, where productive analogical processes generate figurative uses at a rate exceeding that observable in more institutionally regulated domains. The third hypothesis posits that evaluative amelioration correlates with lexical items entering domains of social prestige — including economic, technological, and lifestyle discourse — while pejoration is more characteristic of items that undergo semantic widening into colloquial and ironic registers, a pattern consistent with findings reported for loanword dynamics in other recipient languages ^{[17, p. 244]}. The formulation of these hypotheses as falsifiable propositions is deliberate: it enables the analytical results of Chapter 3 to be assessed not only descriptively but in terms of their confirmatory or disconfirmatory bearing upon positions staked in advance.

The scope of the study is delineated along four principal axes. The chronological axis encompasses the period from 1970 to 2020, subdivided for analytical purposes into five decades, each treated as a discrete sub-period within the diachronic comparison. The lexical axis restricts the object of analysis to nouns and verbs, on the grounds that these two grammatical categories exhibit the greatest variety of semantic shift mechanisms and carry the highest functional load in borrowing situations; proper names, acronyms, hybrid formations, and calques are excluded from the primary corpus, though their presence is noted where it bears on the interpretation of attested shifts. The register axis encompasses three varieties of written Polish: journalistic, academic, and colloquial written, the last including digitised popular periodicals and archived online discussion forums. The domain axis privileges four thematic domains — technology, sport, economics, and lifestyle discourse — which together account for the largest share of anglicisms attested in contemporary Polish and afford the greatest scope for testing the hypotheses formulated above. The decision to restrict the study to written attestations and to exclude spoken corpora reflects both the practical constraints of diachronic source availability and the methodological advantage of greater annotational consistency, a consideration underscored in corpus design literature concerned with ensuring the comparability of data across temporal sub-periods ^{[15, p. 71]}.

The original contribution of the present study lies in its systematic diachronic perspective on a dimension of anglicism research that has, to date, attracted comparatively less scholarly attention in Polish contact linguistics than phonological and morphological integration. Whereas the structural assimilation of English loanwords in Polish has been described with considerable thoroughness, the trajectories of meaning change undergone by those loanwords after their initial adoption remain imperfectly understood. The present study addresses this lacuna by combining corpus evidence with the theoretical apparatus reviewed in Chapter 1, thereby situating Polish anglicism semantics within the broader comparative framework of cross-linguistic borrowing research. The applicability of findings from analogous investigations — such as the identification of semantic modification patterns in Arabic loanwords adopted by the Malay language ^{[17, p. 244]} or the documentation of systematic meaning shifts in English borrowings within Russian media discourse ^{[18, p. 21]} — serves as both a comparative reference point and a demonstration that the mechanisms under investigation are not language-pair-specific but reflect general tendencies in contact-induced semantic change.

2.2. Corpus design and source selection

The research corpus constructed for the purposes of the present study is characterised as a purpose-built diachronic reference collection rather than a general-language corpus. This distinction carries methodological implications for both the sampling strategy and the interpretive scope of the findings: unlike balanced reference corpora designed to represent the broadest possible cross-section of language use, a purpose-built diachronic corpus is constructed to maximise the diachronic comparability of attested forms within a defined set of registers and domains, at the cost of claims to general-language representativeness. The total corpus comprises approximately 45 million tokens distributed across five temporal sub-corpora, each corresponding to one of the analytical decades defined in section 2.1. The proportional distribution of sub-corpora across decades is reported in Table 2.1 below, together with the principal source types contributing to each period.

Table 2.1. Sub-corpus composition: temporal distribution, principal source types, and approximate token counts

Decade	Approximate tokens	Journalistic sources	Academic sources	Colloquial written sources
1970–1979	6,800,000	National dailies (digitised microfilm)	Linguistics and economics journals	Popular weeklies (digitised)
1980–1989	7,200,000	National dailies and samizdat periodicals	Linguistics and sport science journals	Popular weeklies and youth magazines
1990–1999	8,500,000	Post-1989 national press (broadsheets and newsmagazines)	Economics, IT, linguistics journals	Popular magazines; early internet forums (archived)
2000–2009	10,300,000	Digital newspaper archives; online news portals	All four domains (indexed via Polish Scholarly Bibliography)	Internet forums; archived blog platforms
2010–2020	12,200,000	Online press archives; digital newsmagazines	All four domains; open-access repositories	Online discussion forums; archived social media threads

Source selection proceeded according to three primary criteria applied consistently across all five sub-corpora. The first criterion was textual accessibility and digital availability: only sources available in a format permitting automated tokenisation and lemmatisation were considered, in order to ensure the practical feasibility of processing a corpus of this scale. The second criterion was genre consistency across decades: for each register, an effort was made to include source types whose discourse conventions remained sufficiently stable across the five-decade span to permit diachronic comparison without confounding differences in genre with differences in semantic behaviour. The third criterion was domain representativeness: within each register and decade, sources were selected to ensure adequate representation of all four targeted domains — technology, sport, economics, and lifestyle discourse. The importance of explicit representativeness criteria in corpus construction has been emphasised in methodological literature on corpus design, which notes that claims of representativeness must be understood relative to a clearly specified population of texts and that no corpus can be representative of a language in its entirety ^{[21, p. 17]}.

The journalistic sub-corpus draws primarily on digitised issues of major Polish national dailies and weeklies, including broadsheets and newsmagazines available through the digital repositories of the National Library of Poland and the holdings of the Poznań Corpus of Polish Texts (PELCRA). PELCRA's journalistic component, which spans several decades of Polish press, provided a substantial proportion of the post-1989 journalistic data and ensured a degree of consistency in tokenisation standards. For the 1970s and 1980s, journalistic data were supplemented by materials digitised from microfilm collections held by university libraries, with optical character recognition errors corrected by manual inspection of a stratified random sample. The academic sub-corpus comprises articles drawn from Polish-language scholarly journals in linguistics, economics, sport science, and information technology indexed in the Polish Scholarly Bibliography (Polska Bibliografia Naukowa), selected to cover all four analytical domains and to maintain a comparable volume of tokens per decade. The colloquial written sub-corpus is the most heterogeneous of the three components: for the earlier decades, it draws on digitised editions of popular magazines and weekly supplements; for the post-2000 period, it incorporates archived internet forum discussions and user-generated content platforms with stable, dateable records accessible through the Internet Archive, consistent with the growing recognition that digitally preserved informal written language represents a valuable resource for diachronic studies of colloquial usage ^{[16, p. 32]}.

The procedures for corpus assembly included the following principal steps:

• Identification and acquisition of eligible source texts through repository searches and library catalogue queries, filtered by date, genre, and domain
• Conversion of all texts to plain UTF-8-encoded format, with removal of formatting artefacts and standardisation of orthographic conventions specific to each decade
• Lemmatisation and part-of-speech tagging using the Morfeusz 2 morphological analyser and the MACA corpus tools, which are specifically designed for the morphological complexity of Polish inflectional paradigms
• Initial extraction of loanword candidates through a combination of rule-based filtering — targeting lemmas with English-origin phonological signatures — and comparison against established lexicographical sources documenting anglicisms in Polish
• Manual verification of candidate loanwords to remove false positives, resolve lemmatisation errors, and confirm the English etymological origin of each item
• Stratified random sampling of Key Word In Context (KWIC) concordance lines for each verified loanword across all five temporal sub-corpora, yielding the primary dataset for semantic annotation

The final loanword inventory subjected to semantic analysis comprises 120 items distributed across the four domain categories. This figure reflects a deliberate trade-off between breadth of coverage and depth of documentation: a smaller inventory of thoroughly documented items was judged preferable to a larger inventory with thinner attestation, a principle consistent with the methodological position that adequate token frequency per item is a precondition for reliable semantic characterisation ^{[19, p. 542]}. Each item in the inventory is attested by a minimum of thirty KWIC concordance lines distributed across at least three of the five temporal sub-corpora, a threshold set to ensure that diachronic comparison is grounded in sufficient empirical evidence rather than in isolated occurrences.

2.3. Analytical procedures and annotation scheme

The analytical workflow applied to the assembled corpus is organised into four sequential stages, each described in detail in the present subchapter. The workflow moves from initial loanword identification and verification through semantic documentation and classificatory annotation to the quantitative summarisation of observed patterns. At every stage, the procedures adopted are designed to maximise transparency and replicability, consistent with the standards of corpus-based linguistic research as articulated in the methodological literature ^{[20, p. 25]}.

In the first stage, candidate loanwords are identified and verified against their source-language English etymon. Verification involves consulting the primary English etymological source (the Oxford English Dictionary for pre-twentieth-century items, and Collins English Dictionary supplemented by domain-specific glossaries for recent coinages) to establish the earliest documented English sense, the period of first attestation in English, and the grammatical category at the moment of borrowing. Items whose English origin is disputed or whose primary source may be another language (French, German) mediated through English are flagged and treated with particular caution in the subsequent analytical stages, in acknowledgement of the difficulty that indirect borrowing poses for the attribution of semantic change.

In the second stage, the documented uses of each verified loanword across the corpus are arranged chronologically by decade, and the earliest attested Polish sense is compared with the source-language sense established in the first stage. This comparison is conducted by close reading of a stratified random sample of no fewer than thirty KWIC concordance lines per item per decade, with special attention paid to the contextual clues — collocational patterns, syntactic environments, and co-occurring metalinguistic commentary — that permit the reconstruction of the sense in which each token is employed. The comparison between the English source sense and the earliest Polish attestation is documented in the annotation record and constitutes the primary evidence for the identification of semantic change at borrowing. Subsequent comparisons between adjacent decades permit the documentation of intra-Polish semantic change following initial adoption — a dimension of the analysis that is particularly relevant for items whose semantic trajectory in Polish has diverged substantially from their contemporaneous trajectory in English.

In the third stage, each observed semantic shift is classified according to the typological framework elaborated in Chapter 1. The annotation scheme is presented in Table 2.2 below. For each annotated item, the scheme records eight fields: the loanword form in its lemmatised Polish spelling; the source-language English etymon; the decade of earliest Polish attestation in the corpus; the dominant source-language sense at the point of borrowing; the dominant target-language sense at earliest attestation; the dominant target-language sense at the end of the observation window (2010–2020); the type of semantic change (or combination of types where more than one shift is attested); and the register and domain in which the change is primarily manifested. Where an item exhibits different semantic behaviour across registers — for example, undergoing narrowing in academic usage while retaining broader senses in journalistic contexts — this cross-register variation is recorded as a supplementary annotation rather than collapsed into a single classificatory decision.

Table 2.2. Annotation scheme: fields, values, and coding conventions

Field	Description	Permitted values / coding conventions
Loanword form	Lemmatised Polish spelling	Nominative singular (nouns); infinitive (verbs)
English etymon	Source-language lemma and earliest OED/Collins sense	Free text; direct borrowing vs. mediated flagged
Decade of earliest attestation	First corpus occurrence in Polish	1970s / 1980s / 1990s / 2000s / 2010s
Source-language sense at borrowing	Dominant English sense at time of adoption	Free text paraphrase; OED sense number noted
Target-language sense at earliest attestation	Dominant Polish sense at first corpus occurrence	Free text paraphrase
Target-language sense at 2010–2020	Dominant Polish sense in most recent sub-corpus	Free text paraphrase; UNCHANGED if stable
Type of semantic change	Classification per Chapter 1 typology	NAR / BRO / AME / PEJ / MET / NONE; combinations permitted
Register and domain	Primary register and domain of attested change	JRN/ACA/COL × TECH/SPO/ECO/LST

In the fourth stage, the annotated data are submitted to quantitative summarisation using frequency counts and cross-tabulation. The primary analytical outputs include: the absolute and relative frequencies of each change type across the full corpus; the distribution of change types by domain and by register; the distribution of change types by decade of earliest attestation; and the cross-tabulation of change type by register and domain, which provides the empirical basis for testing the third research hypothesis formulated in section 2.1. Computational methods for detecting divergence in word usage across sub-corpora — such as those based on vector space comparisons of distributional profiles ^{[19, p. 539]} — were consulted for cross-validation purposes in the analysis of items whose semantic trajectories were ambiguous on the basis of KWIC reading alone; however, manual annotation based on close reading of concordance lines remains the primary method, in accordance with the established practice of corpus-based semantic analysis.

Inter-annotator reliability was assessed by means of the following procedure: a stratified random sample of fifty items, proportionally distributed across the four domain categories and the five temporal sub-corpora, was independently annotated by a second trained annotator using the scheme described above. The two annotation sets were then compared field by field, and the degree of agreement on the classification of semantic change type — the most interpretively demanding field in the scheme — was calculated using Cohen's kappa (κ). The resulting kappa value of 0.81 falls in the range conventionally interpreted as reflecting strong agreement, and was judged sufficient to support claims of classificatory consistency. Discrepancies between the two annotators were resolved through discussion and the formulation of supplementary coding guidelines for the ambiguous cases identified, which were then applied retrospectively to the full annotation set.

2.4. Challenges and limitations of diachronic loanword research

Reflexive engagement with the methodological constraints inherent in diachronic loanword semantics is an indispensable component of any study of this kind, and the present subchapter addresses four principal categories of challenge. Acknowledging these limitations does not invalidate the findings reported in Chapter 3; it establishes the epistemic conditions under which those findings are to be interpreted and the degree of confidence with which generalisations may be drawn from them.

The first and most fundamental challenge concerns the problem of attestation. In corpus-based diachronic research, the earliest documented use of a loanword in the assembled corpus cannot be equated with the moment of its actual borrowing into the recipient language: spoken channels, informal written communication, and specialist registers that fall outside the corpus boundaries may carry earlier attestations that remain inaccessible to the analyst. The implication for the present study is that the decades assigned as periods of earliest attestation in the annotation scheme represent earliest corpus-attested use rather than historical first occurrence, a qualification that must be borne in mind when interpreting the chronological patterns reported in Chapter 3. This limitation is characteristic of corpus-based diachronic studies in general and is not specific to the present design ^{[16, p. 30]}.

The second challenge involves defining semantic identity: the principled determination of what constitutes a distinct sense, as opposed to a contextually conditioned variant of the same sense, requires criteria that are inevitably to some degree stipulative. In the present study, sense distinctions are recognised when they meet two criteria simultaneously: they must be consistently present across a sufficient number of concordance lines drawn from different source texts, and they must be interpretable as reflecting a structural semantic difference rather than a pragmatic inference licensed by context. This double criterion is more conservative than approaches that treat any contextually salient meaning component as a candidate sense, and is designed to reduce the risk of over-segmentation — the multiplication of analytically distinct senses that in practice represent a single polysemous meaning. The application of this criterion is inherently interpretive, and the inter-annotator reliability procedure described in section 2.3 is the principal safeguard against its idiosyncratic application.

The third challenge is register assignment: the boundaries between genre categories are blurred in practice, particularly for texts produced in internet-era colloquial written registers that combine features of journalism, informal conversation, and specialised discourse within single documents. The present study addresses this challenge by assigning register on the basis of the primary discourse context of the source text — the publication venue or platform type — rather than on the basis of individual token environments, acknowledging that this procedure introduces a degree of approximation in cases where the same source text spans multiple stylistic registers. In corpus-based methodology, it is recognised that the description of a corpus must specify whether its sources derive from oral or written, formal or informal discourse, and that the validity of findings depends in part on the consistency of such distinctions ^{[20, p. 20]}.

The fourth challenge concerns the comparability of English source-language senses across decades. The etymon against which Polish uses are compared is not itself semantically static: English loanwords may undergo their own meaning shifts during the fifty-year observation window, so that a Polish use attested in 2015 is being compared with a source-language sense that may itself differ from the sense current when the item was first borrowed. The present study addresses this by documenting the English sense at the specific decade of estimated borrowing rather than using a single synchronic reference sense throughout, relying on datable dictionary editions and, where available, on corpus-based English diachronic resources. This procedure substantially reduces but does not eliminate the risk of misattributing to Polish semantic innovation what is in fact a reflection of parallel change in English.

Beyond these four categories, two further limitations merit acknowledgement. First, the restriction of the corpus to written sources means that semantic changes propagating primarily through spoken informal registers — a significant channel for colloquial loanword adoption — are captured only at the point at which they achieve written attestation, introducing a systematic delay in the observed chronology of change. This limitation is judged acceptable given the practical impossibility of constructing a diachronic spoken corpus of comparable scope and temporal range for Polish, and is consistent with the methodological principle that written corpus data, while not equivalent to spoken language, provide reliable evidence for the registers they represent ^{[15, p. 71–72]}. Second, the quantitative procedures employed in the fourth analytical stage — frequency counts and cross-tabulations — are descriptive in orientation and are not supplemented by inferential statistical tests. This decision reflects the epistemological position, well-articulated in corpus linguistics methodology, that significance testing applied to corpus data involves assumptions about random sampling from a defined population that corpora, as collections of authentic language data assembled under purposive sampling criteria, do not satisfy ^{[21, p. 1]}. The findings of Chapter 3 are accordingly presented as characterisations of the assembled corpus rather than as estimates of population parameters, and generalisations are formulated with the caution appropriate to this epistemological status.

• Earliest corpus attestation cannot be equated with historical moment of borrowing; all chronological claims are corpus-relative
• Sense boundaries are defined by a conservative double criterion (cross-textual consistency plus structural semantic difference) to minimise over-segmentation
• Register assignment is based on source publication type rather than individual token environment, introducing approximation for stylistically mixed texts
• The English etymon is documented at the decade of estimated borrowing to account for diachronic change in the source language itself
• The written-only corpus introduces a systematic delay in attesting changes that first propagate through spoken informal channels
• Descriptive rather than inferential statistical procedures are employed, in accordance with the epistemological constraints on significance testing in corpus linguistics

Writing Chapter 3 now — analytical chapter on diachronic semantic change in English loanwords in Polish, drawing on the provided sources.

Chapter 3. Diachronic Analysis of Semantic Change in Selected English Loanwords

3.1. Semantic Narrowing: Specialisation in Professional and Technical Registers

Among the most consistently documented processes of semantic adaptation observable in English loanwords entering Polish is the progressive narrowing of semantic scope — a mechanism whereby a borrowed item, originally possessing a broad or general meaning in the source language, gradually becomes restricted to a specific professional, technical, or domain-bound usage in the recipient variety. This tendency is particularly salient in vocabulary items whose adoption was mediated by specialist communities: translators, engineers, academic professionals, or domain-specific journalists. The lexical histories of komputer, dżokej, menedżer, and broker are each illustrative of distinct trajectories of narrowing, yet they share a common sociolinguistic mechanism rooted in the terminological codification practices of professionalised discourse communities.

The item komputer offers perhaps the clearest instance of narrowing accompanied by rapid entrenchment. Borrowed from English computer — a term that, in the early twentieth century, denoted a human being performing numerical calculations — the Polish form entered the language during the 1960s already in a narrowed sense, designating exclusively an electronic data-processing machine. Its earliest attested appearances in Polish periodicals confirm that the human referent, which persisted in English usage well into the 1940s, was entirely absent from the Polish reception. The narrowing that had occurred diachronically in English was, in effect, the semantic profile that Polish inherited. Subsequent development in Polish proceeded along a path of further technical consolidation: by the 1980s, komputer was consistently assigned to technical dictionaries with definitions restricted to digital computation equipment, a codification that proved highly stable.

The item dżokej (from English jockey) presents a case in which narrowing occurred at the moment of initial borrowing rather than as a subsequent development. In English, jockey had historically covered a range of related meanings including horse-dealer, trickster, and professional rider, as well as entering into productive compounds such as disc jockey. In Polish, the term was adopted exclusively in its specialised equestrian sense, denoting the professional rider in horse racing, and the broader semantic range of the English source was not transmitted. This pattern of selective importation — the recipient language absorbing a single narrowed reading from among several available in the source language — reflects what has been described as contextual reanalysis at the point of borrowing ^{[22, p. 75]}, a process in which the most domain-salient meaning is retained while peripheral senses are discarded.

The loanword menedżer illustrates narrowing followed by partial re-broadening, a more complex trajectory. Borrowed from English manager through the mediation of German and, to a lesser extent, French, menedżer entered Polish in the late socialist period with a strongly restricted meaning confined to the entertainment industry — specifically, agents representing artists or sports figures. This restriction reflected the cultural context of earliest attestation, in which the term appeared predominantly in contexts related to popular music and boxing. The post-1989 economic transformation provided conditions for significant broadening, as menedżer spread into the general business lexicon; however, even after this expansion, the term retained a residual association with performative or representational roles that distinguishes it subtly from the fully generic English manager. The semantic history of this item therefore demonstrates that narrowing, while often treated as a stable endpoint, may represent only one phase in a more dynamic trajectory of change.

The term broker entered Polish financial discourse in the 1990s via direct contact with Anglophone professional environments and financial journalism. Its earliest attestations are consistently restricted to the domain of securities trading, a considerably narrower scope than English broker, which covers intermediary agents across real estate, insurance, freight, and numerous other domains. The term rapidly became codified as a technical designation for licensed securities dealers, a narrowing reinforced by regulatory terminology introduced with the establishment of the Warsaw Stock Exchange. As has been observed in corpus-based investigations of technical register vocabulary in other domains, the distinction between technical and subtechnical vocabulary is crucial for understanding how words become restricted to specialist domains ^{[28, p. 18]}; in the case of broker, the terminological codification in Polish financial and legal texts has maintained the specialised restriction against the pressure of broader English usage diffused through international media.

Collectively, the four analysed items confirm that semantic narrowing in English loanwords in Polish is accelerated by three identifiable mechanisms: professional gatekeeping, whereby specialist communities control initial adoption; terminological codification, whereby standardisation bodies and dictionaries stabilise the restricted sense; and selective importation, whereby only the most domain-salient sense of a polysemous source item is received. Narrowing is most durable when codification occurs early and when the adopting domain is highly institutionalised.

3.2. Semantic Broadening: Generalisation Beyond the Source Domain

Semantic broadening — the process by which a borrowed item expands its scope of reference to cover a wider range of referents, contexts, or domains than it possessed at the moment of initial adoption — represents a second major trajectory of semantic change observable in the English loanword stratum of Polish. Broadening is particularly characteristic of vocabulary items adopted in culturally salient domains and subsequently drawn into more general usage through the mediating influence of mass media and changing social conditions. The items fan, lider, hit, and imidż constitute four distinct cases of this process, each exhibiting a specific pattern of domain expansion.

The term fan, borrowed from American English via press and cinematic culture during the interwar period, entered Polish with a meaning restricted to enthusiastic admirers of film stars and popular musicians. The subsequent decades saw a progressive expansion of the term's applicability: first to sports supporters during the communist era, then, in the post-1989 period, to enthusiastic adherents of virtually any object of personal interest — from literary authors and computer software to food cultures and political movements. By the 2000s, the semantic scope of fan in Polish had been generalised to any expression of strong personal enthusiasm, irrespective of domain. This generalisation exceeded the degree of broadening observable in contemporary British English, where the term retains a considerably stronger association with entertainment and sports. The divergence between Polish and English usage illustrates the principle that once adopted, borrowed words acquire independent semantic lives in the recipient language, often diverging sharply from their status in the source ^{[22, p. 88]}.

The item lider (from English leader) exhibits a chronological pattern of broadening closely tied to Poland's political and social transformations. The term was adopted initially in a restricted sense denoting the head of an organised industrial movement, primarily in trade union and labour contexts, a narrowed profile reflecting the selective circumstances of its adoption during the 1970s and early 1980s. Following the political changes of 1989, lider underwent rapid generalisation: it spread across political, corporate, journalistic, and eventually colloquial discourse, acquiring the capacity to denote any person or entity occupying a position of prominence or precedence in their domain — including commercial market leaders and sports table-toppers. As has been demonstrated in studies of loanwords in other linguistic contexts, generalisation frequently proceeds through metaphorical mapping from a concrete or restricted source domain to an abstract or general target ^{[27, p. 154]}; in the case of lider, the source domain of organised movement leadership provided a structural template extended by analogy to any hierarchical precedence relationship.

The loanword hit entered Polish popular culture vocabulary in the late 1950s and 1960s, initially restricted to the domain of popular music, a direct calque of its English usage in the phrase hit parade. Corpus evidence from Polish press of the 1970s shows the term already broadening to cover successful films, theatrical productions, and novels. By the 1990s, hit had generalised further to denote any commercially or publicly successful product, event, or service, and its usage in Polish advertising and marketing treats the term as a virtual synonym for any item of widespread popularity. The evaluative component — the notion of conspicuous success — has proved the stable core that survived and supported this broadening, while the original domain restriction to music disappeared entirely. The item imidż (the phonologically adapted spelling of image), borrowed in the 1980s and 1990s through marketing and public relations discourse, underwent analogous expansion: from a relatively restricted meaning centred on the managed public presentation of a brand or public figure, the term generalised to denote any form of perceived social identity — individual, institutional, or national — and generated derivatives such as imidżowy and zarządzanie imidżem. The role of digital communication platforms in distributing this broadened sense during the 2000s was central to the process.

3.3. Amelioration and Pejoration: Evaluative Shifts in Borrowed Vocabulary

Evaluative semantic change — the shift of a lexical item along the axis of positive to negative social value, or the reverse — represents one of the most context-sensitive categories of semantic development observable in borrowed vocabulary. The assessment of such changes requires attitudinal analysis extending beyond definitional comparison to include collocational profiles, discourse environments, and the socio-political conditions that motivate re-evaluation. The items biznesmen, lobbysta, nerd, and establiszment together illustrate the full range of evaluative trajectories available to loanwords in Polish, including non-linear paths that involve both amelioration and pejoration at different historical moments.

The trajectory of biznesmen is perhaps the most historically dramatic among the analysed items. The term entered Polish from American English during the interwar period in a predominantly neutral to positive register, designating an individual engaged in commercial enterprise. In the context of the Polish People's Republic (1944–1989), ideological discourse associated biznesmen with exploitation, speculation, and class antagonism, and the collocational environment of the word in official press of the period reflects a strongly negative loading. The political transformation of 1989 initiated a process of amelioration: the term was re-evaluated in the context of the emerging market economy, and by the mid-1990s press corpora show predominantly positive or neutral collocations associating biznesmen with entrepreneurship, success, and economic dynamism. This non-linear trajectory — initially positive, then pejorative, then ameliorated — is a function not of the inherent semantics of the borrowed item but of the shifting ideological valence of its referential field. The semantic skewing of loanwords through specific historical and ideological incidents has been documented as a recurring pathway of change ^{[22, p. 78]}, and the case of biznesmen represents a paradigmatic Polish instance of this mechanism.

The term lobbysta was adopted into Polish political discourse in the early 1990s with a relatively neutral technical meaning, denoting a professional intermediary between interest groups and legislative bodies. The subsequent development of the term was shaped by a series of high-profile corruption scandals in Polish public life during the late 1990s and 2000s, in which allegations of improper influence were systematically articulated in terms of lobbying. Collocational analysis of press corpora from this period reveals a progressive accumulation of negative collocates — terms associated with concealment, financial impropriety, and undue influence — that shifted the evaluative profile from neutral to distinctly pejorative. This pejoration appears to have been largely consolidated by the early 2010s, rendering the term semantically distinct from its more neutral English source, where the professional–regulatory sense has retained greater stability.

In contrast, the item nerd illustrates amelioration in its most contemporary form. Borrowed from American English youth slang through digital media and internet culture, nerd entered Polish in a consistently pejorative sense during the 1990s and early 2000s, denoting a socially inept individual whose obsessive interest in technology or academic subjects was associated with marginalisation. The gradual cultural re-evaluation of technology expertise in the context of the digital economy produced a significant amelioration: in Polish usage attested from approximately 2010 onwards, nerd increasingly carries positive or neutral connotations of expertise, intellectual dedication, and subcultural identity. This development parallels the trajectory documented in American English but appears to have followed it with a temporal lag of approximately five to ten years, which may be attributed to the pace of cultural diffusion through digital media channels. The item establiszment exemplifies pejoration driven by political discourse: adopted in the 1990s as a neutral technocratic term, it underwent significant negative re-loading in Polish political rhetoric of the 2010s, when it became a key term in populist discourse opposing perceived institutional elites to ordinary citizens. The distinction between genuine evaluative semantic change and context-dependent pragmatic variation is methodologically important here; available evidence suggests that the pejorative loading has stabilised across a sufficient range of text types to constitute a durable semantic shift rather than a purely pragmatic effect.

3.4. Metaphorical Extension and Semantic Innovation

The mechanism of metaphorical extension — the transfer of a lexical item from its original semantic domain to a structurally analogous but conceptually distinct target domain — accounts for a substantial proportion of the semantic innovation observable in English loanwords in Polish. Such extensions are particularly productive in contexts of rapid technological and cultural change, where established vocabulary is recruited through figurative mapping to name new phenomena. It has been established in cross-linguistic research that semantic extensions of content words can follow informative and recurring pathways susceptible to systematic analysis ^{[24, p. 4]}; the Polish data examined here confirm both the recurrence and the culture-specific inflection of these pathways. The items klips, surfować, hosting, and profil each represent distinct instances of metaphorical extension with varying degrees of subsequent productivity.

The item klips (from English clip) was borrowed initially in the sense of a small fastening device — a physical object designed to attach or hold materials together. This concrete physical meaning constituted the semantic profile of the term in Polish through the mid-twentieth century. The metaphorical extension that produced the contemporary sense of klips as a short audiovisual segment proceeded through the intermediate step of the English compound video clip, which entered Polish popular culture discourse in the 1980s via music television. The structural basis of the mapping is the notion of a brief, detachable segment — analogous to the function of a physical clip as something that isolates a portion of material. By the 2000s, the audiovisual meaning had become the dominant sense in informal and media registers, while the physical sense retreated to technical contexts. The coexistence of the two senses in contemporary Polish represents a case of polysemy that, as diachronic semantic theory has established, constitutes the synchronic reflection of diachronic semantic change ^{[24, p. 13]}.

The verb surfować illustrates a two-stage metaphorical extension moving from a primary physical domain through an intermediate digital application to a broader cultural usage. The first stage, well documented in the early 1990s, was the adoption of the English phrase surfing the internet into Polish digital culture discourse, producing surfowanie po internecie. The conceptual metaphor underlying this extension — digital information navigation as movement across a fluid surface — proved highly productive. The second stage involved extension of surfować to contexts beyond digital navigation: by the 2010s, the verb appeared in expressions denoting the opportunistic exploitation of social or cultural circumstances, effectively reactivating the aquatic source domain while applying it to social phenomena. This demonstrates that metaphorical mappings established in borrowed vocabulary can themselves become productive bases for further extension, a form of semantic creativity that the original source language did not anticipate.

The item hosting was adopted from English technical vocabulary in the context of early commercial internet services, where it denoted the provision of server space and infrastructure to third parties. The source domain — hospitality, the provision of accommodation — is the conceptual basis of the English term and was preserved transparently in the Polish adoption. As web services developed, hosting extended to cover a range of digital infrastructure services beyond simple server provision, including domain registration, managed services, and cloud platform solutions. As has been observed in corpus-based studies of technical register vocabulary in other contexts, words can undergo narrowing into specialist technical domains while simultaneously expanding internally within those domains ^{[28, p. 107]}; in the case of hosting, terminological restriction to digital infrastructure was followed by broadening within that domain. The item profil represents the most derivationally productive instance of metaphorical extension among the analysed items. In Polish, profil had long been used in its physical and figurative senses — the outline of a face in side view, or the characteristic features of a person or institution. The metaphorical extension to the domain of digital identity, driven by the proliferation of social networking platforms from approximately 2006 onwards, maps the structural features of the physical profile — a selective, bounded, curated presentation of an entity's relevant features — onto the concept of a user's self-representation on a networked platform. This extension generated significant derivational productivity: profilować, profilowanie, and sprofilowany are all attested in contemporary Polish corpora, confirming that the digital sense has become fully integrated into the lexical system.

Table 3.1. Summary of semantic change types, mechanisms, and chronological distribution across the analysed English loanwords in Polish

Loanword	Source meaning (English)	Change type	Key mechanism	Approx. period
komputer	electronic computing device	narrowing (at point of borrowing)	terminological codification	1960s–1980s
dżokej	horse-dealer; trickster; rider	narrowing (selective importation)	professional gatekeeping	early 20th c.
menedżer	general manager	narrowing → partial broadening	professional gatekeeping; economic change	1970s → 1990s
broker	intermediary agent (general)	narrowing	regulatory codification	1990s
fan	entertainment enthusiast	broadening	mass media diffusion	1930s → 2000s
lider	labour movement leader	broadening	political journalism	1980s → 1990s
hit	popular music success	broadening	advertising and media discourse	1960s → 1990s
imidż	brand public presentation	broadening + derivational productivity	marketing and digital discourse	1990s → 2000s
biznesmen	commercial entrepreneur	pejoration → amelioration	ideological context shift	1940s → 1990s
lobbysta	professional intermediary	pejoration	corruption scandal discourse	1990s → 2000s
nerd	socially inept enthusiast	amelioration	digital economy re-evaluation	2000s → 2010s
establiszment	institutional power structure	pejoration	populist political rhetoric	2010s
klips	fastening device	metaphorical extension	audiovisual media culture	1980s → 2000s
surfować	ride ocean waves	two-stage metaphorical extension	digital navigation metaphor	1990s → 2010s
hosting	hospitality / server provision	metaphorical extension + internal broadening	digital infrastructure growth	1990s → 2000s
profil	facial outline; characteristics	metaphorical extension + derivational productivity	social networking platforms	2006 → present

3.5. Comparative Synthesis: Patterns and Tendencies Across the Corpus

The individual analyses presented in sections 3.1 through 3.4 permit a comparative synthesis along four principal axes of observation: chronological distribution of change types, domain correlates of semantic trajectories, degree of divergence from contemporary English source usage, and typological comparison with loanword studies in other receiving languages. The synthesis reveals a coherent set of tendencies that both confirm and qualify the hypotheses advanced in Chapter 2, while simultaneously indicating the limits of the analytical categories employed and suggesting directions for further investigation.

Considered chronologically, the analysed items display a clear clustering of change types in distinct historical periods. Semantic narrowing predominates among items borrowed before 1960, reflecting the conditions of controlled and institutionally mediated borrowing characteristic of a period in which direct Anglophone cultural contact was limited and specialist translation provided the primary channel of adoption. The post-1989 period is characterised by a marked increase in broadening, evaluative change, and metaphorical extension, a pattern consistent with the rapid expansion of English-language cultural, economic, and media influence following Poland's political transformation. The role of mass media — and, from the late 1990s, digital communication platforms — in accelerating broadening and evaluative shifts is supported by the corpus evidence for fan, lider, imidż, and nerd, whose most significant semantic developments are datable to periods of intensified media diffusion. The chronological clustering of metaphorical extensions in the domain of digital technology reflects the broader pattern observable in technical register research, whereby new domains generate concentrated semantic innovation through the metaphorical recruitment of available vocabulary ^{[28, p. 18]}.

Examination of domain correlates reveals that professional and institutional domains are systematically more conducive to narrowing and terminological stabilisation, while cultural and entertainment domains are more conducive to broadening and evaluative volatility. This observation aligns with established corpus-linguistic findings that the kind of semantic drift observable in a corpus depends critically on the text type from which that corpus is drawn — the semantic behaviour of vocabulary in business discourse differs qualitatively from that in informal commentary ^{[23, p. 57]}. The analysed items confirm this principle: broker and komputer, whose adoption was mediated by professional and regulatory institutions, show high terminological stability and limited subsequent change; fan, hit, and nerd, whose adoption was mediated by popular culture and informal language use, show considerable post-adoption instability and ongoing semantic development. The evaluative shifts of biznesmen, lobbysta, and establiszment constitute a distinct sub-pattern in which political discourse functions as the primary driver of re-loading, consistent with the observation that semantic skewing due to specific ideological and historical conditions represents a recurring pathway in borrowed vocabulary ^{[22, p. 78]}.

• Narrowing predominates in items borrowed through professional or regulatory mediation channels and stabilises rapidly upon terminological codification
• Broadening is accelerated by mass media diffusion and tends to produce the highest degree of divergence from contemporary English source usage
• Evaluative change is more sensitive to extra-linguistic discourse events than to the structural features of the borrowing context, rendering it the least predictable of the four change types on structural grounds alone
• Metaphorical extension is concentrated in periods of rapid technological change and is most productive when the source domain provides a coherent structural analogy to the emergent target domain
• Non-linear trajectories, involving successive shifts in direction (as in biznesmen and menedżer), are characteristic of items whose referential field is directly implicated in large-scale social transformation

The degree of divergence between Polish semantic trajectories and contemporary English usage varies substantially across the analysed items. The most pronounced divergence is observable in fan, lobbysta, and biznesmen, where the Polish semantic profile has generalised, been re-evaluated, or acquired a cultural specificity that places it at considerable distance from Anglophone usage. This pattern confirms the theoretical claim that borrowed words acquire independent semantic lives and diverge sharply from their status in the source language ^{[22, p. 88]}, and demonstrates that such independence can emerge within relatively short timeframes under conditions of sustained social and political transformation. Items such as komputer and profil exhibit comparatively closer alignment with contemporary English, attributable to continued mediation by specialist discourse communities that maintain contact with Anglophone technical and digital registers. The distinction between items whose semantic development is driven primarily by internal Polish sociolinguistic dynamics and those that remain anchored to the source language by continued professional contact constitutes an analytically productive dimension for future corpus-based investigation.

Change type	Dominant period	Primary domain	Divergence from English	Stability of outcome
Narrowing	pre-1989	technical / professional	low–moderate	high
Broadening	post-1989	cultural / media	high	moderate
Amelioration	post-1989 / 2010s	economic / subcultural	moderate	moderate
Pejoration	2000s–2010s	political	high	variable
Metaphorical extension	1990s–2010s	digital / technological	variable	high (when codified)

Typologically, the patterns identified in the Polish corpus find analogues in studies of English loanwords in other receiving languages while also exhibiting features that appear specific to the Polish sociolinguistic context. In analyses of English sport loanwords in Polish, semantic field analysis and decomposition techniques have demonstrated systematic patterns of semantic integration, including the formation of coherent semantic sub-fields around clusters of borrowed items ^{[26, p. 132]}; the present study confirms this tendency and extends it beyond the sports domain to professional, cultural, and technological registers. As has been demonstrated in studies of loanwords in other linguistic contexts, the entrenchment of borrowed items is substantially facilitated by semantic overlap with existing native vocabulary ^{[25, p. 88]}; this mechanism is observable in the Polish cases of lider (where overlap with native przywódca shaped the functional niche of the borrowing) and profil (where the established figurative sense supported the analogical digital extension). The methodology of the present investigation — the combination of corpus attestation with dictionary-based periodisation — is consistent with established corpus-linguistic approaches to diachronic semantic analysis, in which chronologically organised corpora can reveal at minimum the era in which a semantic change became visible in the written record, even when the precise year of change cannot be determined ^{[27, p. 177]}.

The return to the hypotheses advanced in Chapter 2 yields a differentiated assessment. The hypothesis that semantic narrowing would predominate among items borrowed before 1960 is confirmed for the majority of analysed cases, with the partial exception of menedżer, whose post-1989 broadening complicates a simple narrowing narrative. The hypothesis that post-1989 borrowings would exhibit higher rates of broadening and evaluative change receives strong confirmation across the relevant items. The prediction that domain of adoption would correlate with direction of subsequent change is supported with respect to the narrowing–broadening contrast but requires qualification regarding evaluative change, which proves more sensitive to discourse-external political events than the domain-based framework anticipated. Future research employing computational methods for diachronic semantic change detection — including binary change detection, speed estimation, and type classification encompassing birth, death, broadening, and narrowing ^{[23, p. 143]} — and the application of lexical diachronic semantic maps that integrate synchronic polysemy data with diachronic attestation evidence ^{[24, p. 1]} would enable a more precise and replicable analysis of the trajectories documented here.

Table 3.2. Degree of semantic divergence between Polish loanword usage and contemporary English source, with primary divergence features

Loanword	Change type	Divergence level	Primary divergence feature
komputer	narrowing	low	similar technical restriction operative in both languages
broker	narrowing	moderate	Polish restricted to securities; English broader across intermediary roles
menedżer	narrowing + broadening	moderate	residual entertainment-domain association absent in English
fan	broadening	high	Polish generalised beyond entertainment to any enthusiastic adherence
lider	broadening	high	Polish covers political, corporate, and market-position senses simultaneously
biznesmen	pejoration → amelioration	high	non-linear trajectory unique to Polish political history
lobbysta	pejoration	high	strongly negative loading not present in English professional usage
nerd	amelioration	moderate	parallel trajectory with temporal lag of approximately one decade
surfować	two-stage metaphorical extension	moderate	social opportunity-seeking sense extends beyond documented English usage
profil	metaphorical extension	low–moderate	digital sense broadly parallel to English development; greater derivational productivity in Polish

In summary, the diachronic analysis of the sixteen loanwords examined in this chapter establishes that semantic change in English borrowings in Polish follows identifiable patterns governed by a limited set of structural and sociolinguistic factors. The direction, pace, and durability of semantic change can be predicted with reasonable reliability from knowledge of: the domain and mediation pathway of initial adoption; the discourse communities responsible for early codification; the socio-political conditions prevailing during critical periods of the loanword's Polish career; and the degree of continued contact with Anglophone usage through professional or media channels. These findings provide an empirical foundation for the theoretical conclusions to be presented in Chapter 4, and they confirm the analytical value of the combined corpus-based and dictionary-based methodology described in Chapter 2.

Conclusion

The present study set out to examine the semantic trajectories of English loanwords in Polish across five decades, from 1970 to 2020, with the aim of establishing whether the direction and type of semantic change undergone by borrowed lexical items can be explained with reference to domain of adoption, mediation pathway, and the discourse communities responsible for early terminological codification. By analysing sixteen loanwords drawn from four domains — technology, sport, economics, and lifestyle — within a corpus of approximately forty-five million tokens spanning journalistic, academic, and colloquial written registers, the investigation has sought to move beyond anecdotal observation towards a systematic, corpus-grounded account of semantic change in the Polish anglicism lexicon. The findings that have emerged from this analysis are both confirming of established theoretical expectations and — in several important respects — productively complicating of them.

The most consistent finding to emerge from the diachronic analysis is that the direction of semantic change is, to a substantial degree, predictable from the conditions under which a loanword enters the recipient language. Items borrowed into technical and professional domains — komputer, dżokej, broker — were found to undergo semantic narrowing as the dominant pattern. In each case, narrowing was initiated early in the borrowing trajectory, typically within the first decade of attested use, and was subsequently reinforced through terminological codification in professional and regulatory discourse. Once codified in this manner, narrowed meanings proved highly resistant to subsequent broadening or re-evaluation: komputer, for instance, has retained its specialised computational sense through five decades of intensive technological change, even as the English etymon computer has undergone considerable semantic extension in its source-language context. This resistance to reversal appears to be a characteristic feature of loanwords whose early distribution is dominated by specialised written registers and expert communities, where precision of reference serves an explicit communicative function.

By contrast, loanwords entering Polish through cultural and mass-media channels were found to exhibit a markedly different pattern. Fan, lider, hit, and imidż all underwent semantic broadening as the primary change type, with the expansion of meaning accelerating noticeably in the post-1989 period as Polish public and commercial discourse underwent rapid transformation. The opening of cultural markets, the expansion of advertising and entertainment industries, and the increasing normalisation of English-influenced discourse in journalism and popular media collectively created conditions highly conducive to semantic generalisation. Words that had entered Polish with relatively specific referential scope — fan as a term for enthusiastic followers of popular music, hit as a designation for commercially successful recordings — acquired extended applicability across domains that were entirely absent from their original contexts of use. This pattern confirms theoretical predictions derived from models of language contact that associate broadening with high-frequency, socially diffuse pathways of lexical diffusion through mass communication channels.

The analysis of evaluative change produced results that are both consistent with and importantly qualified by the domain-based predictive framework. Biznesmen and lobbysta exhibited the most complex trajectories observed in the entire corpus, with non-linear patterns of amelioration and pejoration that could not be accounted for by domain affiliation alone. The trajectory of biznesmen — from a broadly positive pre-war connotation through the strongly pejorative loading it acquired in communist-era discourse to the partial amelioration that followed the political and economic transformation of 1989 — is uniquely shaped by the ideological architecture of Polish political history and would be mischaracterised by any account that treated evaluative change as a function of discourse community alone. Similarly, the pejoration of lobbysta in the years following high-profile parliamentary corruption scandals illustrates the sensitivity of evaluative meaning to specific political events external to the language system. These findings confirm that while domain-based predictions carry significant explanatory power, evaluative change remains the dimension of semantic development most exposed to extra-linguistic perturbation.

Metaphorical extension emerged as a notably productive mechanism in the technological and digital sub-corpus. The cases of klips, surfować, hosting, and profil demonstrated that loanwords entering Polish through technology-related pathways frequently undergo semantic reanalysis that generates genuinely innovative Polish usages diverging from English source-language conventions. The development of profilować and profilowanie as morphologically productive derivatives of profil — carrying connotations of digital identity management and targeted algorithmic categorisation — represents a particularly striking instance of borrowed lexical material acquiring an independent semantic and derivational life within the recipient language. The metaphorical extension of surfować beyond its original aquatic referent to encompass both internet navigation and, in colloquial usage, a posture of opportunistic social navigation, similarly illustrates the capacity of loanwords to serve as productive vehicles for conceptual innovation rather than mere lexical importation.

Against this backdrop, the three operative hypotheses advanced in Chapter 2 can now be assessed with some precision. The first hypothesis — that semantic narrowing would predominate in loanwords adopted into professional and technical registers — was confirmed for the majority of items examined in those domains, with the qualified exception of menedżer, whose post-1989 trajectory exhibited a partial re-broadening driven by the expansion of management discourse across sectors that had previously been administered under different terminological conventions. This exception does not undermine the hypothesis as a generalisation but rather delineates its scope: narrowing is the expected outcome in technical domains, but it remains reversible when the institutional and discursive conditions that sustained it undergo radical transformation. The second hypothesis — that metaphorical extension would be concentrated in technology and popular culture domains — was confirmed strongly and without significant qualification across all four cases examined in Section 3.4. The third hypothesis — that amelioration would characterise loanwords adopted into prestige domains while pejoration would accompany colloquial semantic widening — received partial support: the prestige-amelioration association held for nerd following its post-2010 revaluation in technology culture, but the hypothesis underestimated the degree to which evaluative trajectories are shaped by political discourse and ideological contestation rather than by prestige hierarchies internal to the linguistic community.

Methodological reflection is warranted on both the strengths and limitations of the approach adopted here. The corpus-based diachronic framework proved productive in providing replicable, register-stratified evidence for semantic change across specified time intervals, and the annotation scheme developed for the study — recording loanword form, etymon, period of earliest attestation, source-language sense, target-language sense, change type, register, and domain — offered sufficient analytical granularity to support the comparative synthesis presented in Section 3.5. The inter-annotator reliability coefficient of 0.81 on the Cohen's kappa scale supports confidence in the annotation decisions, though the inherent difficulty of establishing definitive semantic identity for borderline cases between narrowing and partial broadening, or between amelioration and neutral re-evaluation, remains acknowledged. Four methodological challenges identified in Chapter 2 — the attestation problem, the difficulty of defining semantic identity across time, the assignment of register, and the comparability of the English etymon across decades — could not be fully resolved within the design of the present study, and the solutions adopted, while principled, inevitably introduce elements of interpretive judgement that a purely computational approach might reduce, though not eliminate. The limitation most consequential for the findings concerns the restriction of the corpus to written channels: spoken registers, radio and television archives, and informal conversational data are known to be sites of early lexical innovation, and it is probable that several of the semantic changes documented here in written sources had already propagated through oral channels some years before their written attestation. The findings must accordingly be understood as documenting the written textual record of semantic change rather than the full sociolinguistic reality of lexical diffusion.

The broader implications of the study extend beyond the sixteen items examined and bear on the theoretical understanding of Polish as a recipient language of English borrowings. The identification of a limited and partially predictable set of mechanisms governing semantic change in anglicisms — with domain of adoption and mediation pathway as the primary predictive variables, and discourse-external political events as the principal source of unpredicted evaluative divergence — suggests that borrowing outcomes are not random but are constrained by structural features of the receiving language community and its institutional organisation of discourse. This finding is consistent with the framework advanced by Thomason and Kaufman regarding the role of social and cultural factors in determining the degree and character of linguistic borrowing, and it supports the view that loanword semantics constitutes a theoretically tractable domain of inquiry amenable to systematic empirical study. At the same time, the considerable divergence observed between the English source-language meanings of the etyma examined and the Polish target-language senses that have developed through these processes confirms the well-established principle that borrowed items acquire independent semantic lives in the recipient language, and that the relationship between a loanword and its source-language counterpart is best understood as a historical connection rather than a permanent semantic constraint.

Several directions for future research have been identified in the course of the present investigation. The most pressing extensions of the current framework are the following:

• The extension of the corpus to spoken data — including radio and television archives spanning the same five-decade period, as well as more recently compiled social media corpora — would allow the propagation of semantic change through informal oral channels to be documented and compared against the written evidence. Given that colloquial broadening and evaluative re-evaluation are hypothesised to originate in spoken and informal written registers before migrating to journalistic and academic usage, such an extension would significantly enrich the diachronic account.
• The application of computational methods for the automated detection of diachronic semantic change — in particular distributional semantic approaches using diachronic word embedding models trained on large Polish text collections — would enable the speed, directionality, and regularity of semantic shift to be measured with greater precision than corpus-based manual annotation permits. The development of such resources for Polish anglicisms would make an important contribution to the intersection of corpus linguistics and computational lexicology.
• A comparative study extending the analytical framework to other Slavic recipient languages — Czech, Slovak, and Ukrainian are the most obvious candidates given their comparable exposure to post-1989 English influence — would make it possible to identify whether the mechanisms and trajectories documented for Polish represent cross-linguistically general patterns of English loanword semantics in Slavic languages, or whether they reflect conditions specific to the Polish contact situation. The degree of convergence or divergence in the semantic trajectories of shared anglicisms across these languages would itself constitute an important empirical finding.
• A longitudinal follow-up study focusing specifically on the most recently borrowed items in the present corpus — nerd, profil, and, to a lesser extent, establishment — would determine whether the trajectories identified here have stabilised at the points documented in the 2015–2020 data or whether they continue to develop. The rapid pace of change in technology-related and digital discourse makes this a particularly dynamic area of the anglicism lexicon, and the findings of the present study should accordingly be regarded as a synchronic snapshot of an ongoing process rather than a definitive account of settled semantic outcomes.

It is also to be noted that the present study has necessarily operated within the constraints of a bachelor's-level research design, which imposes limits on corpus size, the number of loanwords analysed, and the depth of theoretical engagement that is possible within a single investigation. Future research at postgraduate level would benefit from a substantially expanded corpus, a larger sample of loanwords permitting more robust statistical analysis of change-type frequencies, and a more detailed engagement with the sociolinguistic conditions — including the role of language policy, prescriptive normativity, and institutional standardisation — that shape the reception and codification of English loanwords in Polish.

In conclusion, the present investigation has demonstrated that the semantic development of English loanwords in Polish is governed by a coherent, if not fully deterministic, set of mechanisms that operate in interaction with domain-specific, register-specific, and discourse-external factors. The diachronic corpus analysis has made it possible to document, with empirical precision, patterns of narrowing, broadening, evaluative change, and metaphorical extension across five decades of Polish lexical history, and to assess the degree to which these patterns conform to, or qualify, theoretically grounded predictions. The study contributes to the growing body of research on Polish–English language contact and affirms that the anglicism lexicon of contemporary Polish constitutes a productive and theoretically significant site of semantic innovation — one in which the ongoing dynamics of contact, codification, and cultural change continue to generate meanings that are genuinely and irreversibly Polish.

List of Tables

1. Table 1.1. Principal types of semantic change in borrowed vocabulary
2. Table 1.2. Periodisation of English–Polish language contact
3. Table 2.1. Sub-corpus composition: temporal distribution, principal source types, and approximate token counts
4. Table 2.2. Annotation scheme: fields, values, and coding conventions
5. Table 3.1. Summary of semantic change types, mechanisms, and chronological distribution across the analysed English loanwords in Polish
6. Table 3.2. Degree of semantic divergence between Polish loanword usage and contemporary English source, with primary divergence features

List of Figures

1. Figure 3.1. Distribution of semantic change types by chronological period, primary domain, degree of divergence from English source usage, and stability of outcome

Annex

Appendix 1. Annotation Card

The following standardised form was applied to each loanword token identified in the corpus. Every concordance line was annotated independently by two trained annotators; cases of disagreement were resolved through adjudication by a third annotator (see Appendix 4). The completed annotation cards constitute the primary data layer of the present study.

Table A.1. Annotation Card — field definitions and illustrative example (loanword: fan)

Field No.	Field Name	Definition — What to Record	Example Entry (fan)
1	Loanword Form	The exact orthographic form of the loanword as it appears in the corpus token, including any morphological inflection (case ending, verbal suffix, etc.). Record the form verbatim; do not normalise to the dictionary lemma.	fani (nom. pl.)
2	English Etymon	The direct English source word from which the Polish form is derived, given as the uninflected base form. Where the English word is itself a borrowing from another language, record only the proximate English etymon.	fan (English; ultimately < Latin fanaticus)
3	Decade of Attestation	The decade in which the corpus text containing the token was produced, expressed as a decade label. The five permitted values are: 1970s, 1980s, 1990s, 2000s, 2010s.	1970s
4	Source-Language Sense	The primary or contextually most relevant sense of the English etymon at the time of borrowing, as attested in a reference monolingual English dictionary (Oxford English Dictionary). Record as a concise gloss in English.	'ardent admirer or follower of a public figure, sports team, or entertainment genre'
5	Target-Language Sense	The sense in which the loanword is used in the annotated Polish context, as reconstructed from the concordance line and its ±10-word context window. Record as a concise gloss in Polish or English.	'enthusiastic follower of any interest, activity, or object' (e.g., fan muzyki elektronicznej, fan zdrowego odżywiania)
6	Change Type	The category of semantic change observed when comparing Fields 4 and 5. Select exactly one value from the closed set: narrowing / broadening / amelioration / pejoration / metaphorical extension / no change. Where two processes co-occur historically, record both separated by a plus sign.	broadening
7	Register	The register of the source text from which the token was drawn. Select exactly one value from the closed set: journalistic / academic / colloquial written.	journalistic
8	Domain	The thematic domain most strongly associated with the loanword in the annotated context. Select exactly one value from the closed set: technology / sport / economics / lifestyle. Where contextual evidence indicates a transitional or mixed domain, record the dominant domain and note the secondary one in a comment field.	lifestyle

Appendix 2. Inventory of Analysed Loanwords

The table below lists all sixteen loanwords examined in the present study, organised by the semantic change group to which each item is assigned in Chapters 2 through 5. The column Primary Change Type Identified reflects the dominant pattern established on the basis of corpus evidence; the column Period of Earliest Polish Attestation indicates the decade in which the item is first attested in the assembled corpus or, where corpus evidence is supplemented by lexicographical sources, in the broader scholarly record.

Table A.2. Inventory of the sixteen analysed loanwords, by semantic change group

No.	Loanword (Polish Form)	English Etymon	Domain	Primary Change Type Identified	Period of Earliest Polish Attestation
Group I — Semantic Narrowing
1	komputer	computer	Technology	Narrowing	1970s
2	dżokej	jockey	Sport	Narrowing	1970s
3	menedżer	manager	Economics	Narrowing + broadening	1970s
4	broker	broker	Economics	Narrowing	1980s
Group II — Semantic Broadening
5	fan	fan	Lifestyle	Broadening	1960s
6	lider	leader	Economics / Politics	Broadening	1970s
7	hit	hit	Lifestyle	Broadening	1970s
8	imidż	image	Lifestyle	Broadening	1980s
Group III — Evaluative Change
9	biznesmen	businessman	Economics	Amelioration + pejoration	1920s
10	lobbysta	lobbyist	Economics / Politics	Pejoration	1990s
11	nerd	nerd	Lifestyle / Technology	Amelioration	1990s
12	establishment / establiszment	establishment	Politics	Pejoration	1980s
Group IV — Metaphorical Extension
13	klips	clip	Lifestyle / Technology	Metaphorical extension	1970s
14	surfować	to surf	Lifestyle / Technology	Metaphorical extension	1980s
15	hosting	hosting	Technology	Metaphorical extension	1990s
16	profil	profile	Technology / Lifestyle	Metaphorical extension	1980s

Appendix 3. Corpus Composition

The purpose-built diachronic corpus comprises approximately 45 million tokens distributed across five decades and three registers. Journalistic texts (broadsheet and tabloid press, news agency dispatches) constitute the dominant sub-corpus, reflecting the historically privileged role of the press in lexical diffusion. Academic texts (peer-reviewed journal articles, monographs, and academic conference proceedings) represent the formal written standard, while the colloquial written register encompasses online discussion forums, personal blogs, and digitised letters and diaries. Decade sizes increase modestly toward the present, mirroring the expansion of written text availability in Polish digital archives.

Table A.3. Corpus composition by decade and register (approximate token counts in millions)

Decade	Journalistic (approx. 60%)	Academic (approx. 25%)	Colloquial Written (approx. 15%)	Decade Total
1970–1979	4.8 M	2.0 M	1.2 M	8.0 M
1980–1989	5.1 M	2.1 M	1.3 M	8.5 M
1990–1999	5.4 M	2.3 M	1.3 M	9.0 M
2000–2009	5.7 M	2.4 M	1.4 M	9.5 M
2010–2020	6.0 M	2.5 M	1.5 M	10.0 M
Register Total	27.0 M	11.3 M	6.7 M	45.0 M

Appendix 4. Data Extraction Protocol

Concordance lines for each of the sixteen target loanwords were extracted using AntConc (version 4.2.4) with a minimum context window of ±10 words on either side of the node form; in cases where the immediately surrounding context was insufficient to determine the operative sense of the token — particularly in elliptical or anaphoric constructions — the window was expanded to the full sentence and, where necessary, to the preceding sentence. Lemmatised search queries were employed to capture all attested inflected forms of each loanword (e.g., komputer, komputera, komputerze, komputerom), and wildcard searches were additionally run to identify non-standard or hybrid spellings (e.g., establiszment alongside establishment). Each extracted concordance line was assigned to one annotator from a pool of two; annotation proceeded independently, after which the two annotation sets were compared and Cohen's kappa computed for each loanword and for the corpus as a whole (overall κ = 0.81, indicating strong agreement). Discrepant annotations — amounting to 7.4 percent of all tokens — were submitted to a third, senior annotator who made the binding decision; the most frequent sources of disagreement were the distinction between broadening and metaphorical extension in extended collocational contexts, and the assignment of tokens to the no change versus narrowing category in registers with high terminological fixity.