Fact-checked by Grok 2 weeks ago

Wave model

The wave model, also known as the wave theory or Wellentheorie, is a foundational concept in that describes how linguistic innovations—such as new sounds, words, or grammatical features—spread geographically from a point of origin in a manner resembling concentric ripples on water, leading to overlapping and intersecting patterns of change across dialects or languages. This model emphasizes diffusion through ongoing contact between speakers rather than abrupt separations, using isoglosses (boundaries marking the spread of specific innovations) to map gradual transitions in linguistic features. Proposed by German linguist Johannes Schmidt in his 1872 work Die Verwandtschaftsverhältnisse der indogermanischen Sprachen, the wave model emerged as a direct critique of August Schleicher's earlier family tree (Stammbaum) model, which depicted language evolution as rigid, branching divergences without accounting for lateral influences. Schmidt argued that formed through waves of innovation radiating from cultural or political centers, resulting in bundles of isoglosses that could overlap and create dialect continua rather than strictly nested subgroups. This approach highlighted the role of areal diffusion in , particularly in regions with high speaker interaction, such as ancient or modern . In contrast to the family tree model's assumption of sudden splits and isolation, the wave model accommodates ambiguity in linguistic relationships, where a single language or dialect might participate in multiple partially overlapping innovation waves, better explaining phenomena like Sprachbünde (linguistic areas) or continuum languages. Early support came from dialect geographers like Jules Gilliéron, whose work on Romance dialects in the early 20th century provided empirical evidence through detailed isogloss mapping. The wave model has evolved significantly in modern , influencing dynamic extensions by Charles-James N. in the , which incorporated implicational hierarchies to sequence innovations temporally, and Malcolm Ross's linkage model in 1988, which applies wave principles to family-level diversification in cases like . Recent quantitative methods, such as historical glottometry developed by Siva Kalyan and Alexandre François in 2018, use computational analysis of hundreds of innovations to measure bundling strength, validating the model's predictions in diverse language families. Today, it remains a key tool for understanding hybrid language histories, complementing phylogenetic approaches in fields like computational .

Background and History

Origins in 19th-Century Linguistics

In the mid-19th century, proposed the Stammbaumtheorie, or family tree model, as a framework for understanding linguistic relationships, particularly within the . This model depicted languages diverging through strict bifurcating from a common ancestor, akin to branches on a , and explicitly excluded horizontal influences such as borrowing or diffusion between related languages. Schleicher illustrated this concept in works like his 1853 essay and 1863 diagram of , emphasizing vertical descent without accounting for areal contacts. This rigid approach faced criticism for failing to explain observed patterns of gradual, overlapping changes in dialects, prompting a shift toward alternative models. In 1872, Johannes Schmidt articulated the wave model (Wellentheorie) in his publication Die Verwandtschaftsverhältnisse der indogermanischen Sprachen, proposing that linguistic innovations spread like ripples from concentric circles originating at a , diffusing unevenly across a geographic rather than following branches. Schmidt's theory highlighted how sound changes and other features could propagate horizontally through neighboring varieties, blending influences in ways incompatible with the family tree's isolationist assumptions. Early support for Schmidt's ideas came from Schuchardt, who in the 1870s advocated for the based on his studies of , arguing that their gradual, blending evolutions—such as vowel shifts and morphological mergers—demonstrated continuous intermixture rather than sharp bifurcations. Schuchardt critiqued the for overlooking these fluid transitions, using examples from Iberian and Gascon dialects to show how innovations permeated borders organically. A classic illustration of the wave model is the (Zweite Lautverschiebung), where stops like /p/, /t/, and /k/ fricativized or affricated unevenly across southern German dialects, forming isoglosses that bundled irregularly rather than aligning with tree-like divisions between Low and . This diffusion pattern underscores the model's emphasis on spatial gradients over strict genealogical splits.

Key Developments and Proponents

The wave model, initially proposed by Johannes Schmidt in his 1872 work Die Verwandtschaftsverhältnisse der indogermanischen Sprachen, underwent significant expansions in the late 19th century through the critiques and theoretical refinements of key linguists challenging the rigid family-tree model of the Neogrammarians. Hugo Schuchardt played a pivotal role in this development, particularly in his 1885 pamphlet Über die Lautgesetze: Gegen die Junggrammatiker, where he critiqued the Neogrammarian insistence on exceptionless sound laws and advocated for a more fluid, wave-like blending of linguistic innovations across dialect borders, emphasizing the concept of Bund (bundle or league) to describe gradual transitions rather than sharp divisions. Schuchardt's ideas highlighted how language change occurs through social diffusion, influencing subsequent dialectological studies by underscoring the porous nature of linguistic boundaries. In the early 20th century, empirical validation of wave theory advanced through the work of dialect geographers, most notably Georg Wenker's pioneering Sprachatlas des Deutschen Reichs (), which mapped isoglosses for features like vowel shifts across thousands of German localities, visually demonstrating wave patterns of innovation spread and bundle-like continua. Wenker's atlas, based on questionnaires distributed to over 50,000 sites, provided the first large-scale cartographic evidence for overlapping waves of change, shifting focus from abstract theory to observable spatial distributions and inspiring similar projects in other languages. Mid-20th-century further integrated concepts of and akin to the wave model into studies of ongoing change, with William Labov's research—such as his analysis of in —showing how innovations spread unevenly through communities based on social factors like prestige and contact. Labov's empirical approach, combining quantitative variation analysis with models of change , bridged 19th-century and modern , emphasizing how social dynamics facilitate the spread of features like shifts. By the early 2000s, the wave model saw formal theoretical in the work of Malcolm Ross, who introduced the "linkage" concept in papers on Austronesian languages, such as his study on contact-induced change in north-west , to describe non-tree structures where dialects form interconnected groups rather than bifurcating branches. Ross's framework, elaborated in subsequent works up to 2007, formalized wave theory for complex families by positing "linkages" as networks of communalects linked by shared innovations across sprachbunds, avoiding the limitations of strict genealogical trees. In applying this to , Ross identified numerous linkage groups, where linguistic features spread laterally through areal contact rather than vertical descent, providing a robust model for reconstructing histories in regions with high mixing. This marked a culmination of wave theory's evolution from Schmidt's proposal, incorporating empirical and sociolinguistic insights to address diverse language diversification patterns.

Core Principles

Mechanism of Innovation Spread

In the wave model, linguistic innovations—such as phonological shifts or lexical borrowings—originate at a specific point within a and propagate outward in a manner analogous to ripples expanding across a surface. This occurs through ongoing contact among speakers of mutually intelligible varieties, influencing adjacent dialects to varying degrees depending on the duration and intensity of interaction over time. Johannes Schmidt illustrated this process with a concentric circle diagram, depicting the innovation's epicenter where it first emerges, followed by rings of decreasing intensity as the feature spreads to peripheral areas. These expanding waves often intersect, resulting in isoglosses—geographic boundaries demarcating the presence or absence of the innovation—which may bundle multiple overlapping waves to form dialect boundaries. Such visualization underscores how innovations do not replace entire languages uniformly but permeate dialects selectively, creating a of partial adoptions. The spread of an innovation is modulated by several key factors, including geographic and social proximity between speakers, the prestige associated with the originating group, and resistance from entrenched linguistic features in target dialects. For instance, i-umlaut in , which involves vowel fronting triggered by a following high front , originated in early varieties around the 6th century and diffused unevenly southward, appearing regularly in and but with partial or delayed implementation in continental West Germanic dialects due to varying degrees of contact and phonological resistance. Unlike the family tree model, which posits discrete bifurcations and uniform inheritance down branches, the wave model rejects assumptions of isolated evolution, instead accommodating partial overlaps, retreats of innovations, and retractions where waves fail to penetrate fully. A critical element is the wave front, the advancing boundary of maximum effectiveness where the innovation achieves peak diffusion before attenuating, often yielding stepped rather than smooth gradients in the resulting dialect continuum.

Dialect Continuum and Wave Propagation

A refers to a geographical of linguistic varieties where adjacent dialects exhibit minimal differences and remain mutually intelligible, yet cumulative variations across the chain result in significant divergence between distant endpoints. This structure arises from ongoing horizontal diffusion of linguistic features rather than vertical from a common ancestor, producing a seamless of change without boundaries. A prominent example is the dialect continuum, spanning from in the west to in the east, where neighboring varieties share high intelligibility, but peripheral forms like Maghrebi and differ markedly in , vocabulary, and syntax. In the wave model, the propagation of multiple innovations creates overlapping diffusion patterns, leading to bundled isoglosses—concentrations of linguistic boundaries—that form broad transition zones known as in German dialectology, rather than the sharp bifurcations of family tree models. These zones represent areas of intensified contact and mixing, where features from various waves intersect and blend, fostering gradual shifts across the continuum. Johannes Schmidt's 1872 diagram illustrates this process through intersecting concentric circles emanating from innovation foci, depicting how waves generate "stepped" patterns of change that accumulate into networks of interconnection, without hierarchical splits. Central to this framework is the absence of a single "parent" language; instead, observed variations emerge from the cumulative effects of horizontal transmissions, where innovations spread laterally through adjacent communities over time. Such processes underscore the wave model's emphasis on contact-induced evolution, as seen in the , where retroflex consonants—absent in early Indo-European—spread via contact with substrate languages in southern contact zones, creating a wave-like pattern across the subcontinent.

Applications

In Dialectology and Areal Linguistics

The wave model has been instrumental in for empirically mapping the spatial distribution of linguistic features, as demonstrated by Georg Wenker's pioneering survey of conducted between 1876 and 1887, which involved distributing questionnaires with 40 test sentences to over 50,000 localities to identify variations in , , and syntax. This effort produced the first comprehensive dialect atlas, revealing bundles of isoglosses—lines demarcating areas where specific linguistic traits predominate—that aligned with the model's prediction of innovations spreading radially like waves from focal points, rather than forming strict tree-like boundaries. Wenker's work provided early quantitative validation for wave propagation in real-world dialect continua, influencing subsequent atlases by emphasizing gradual transitions over abrupt divisions. In areal linguistics, the wave model elucidates how linguistic features diffuse across genetically unrelated languages within a , such as the Balkan linguistic area, where shared traits like postposed definite articles (e.g., Romanian omul 'the man' from om-ul) and evidential verb forms have spread through prolonged contact among , , , and , independent of common ancestry. This diffusion exemplifies wave-like propagation, with innovations emanating from cultural or trade hubs and attenuating over distance, fostering convergence in a continuum that defies family-tree classifications. Such patterns highlight the model's utility in tracing contact-induced changes, as seen in the Balkan region's postposed articles, which likely originated in Romance or and radiated eastward via multilingual interactions. Modern dialectometry builds on the wave model by employing quantitative metrics to measure linguistic distances and visualize gradients, such as the Levenshtein adapted for , which calculates the minimum operations needed to transform one dialectal into another, thereby modeling wave-like smoothing in continua. Jean Séguy's 1971 coefficient, applied to Romance dialects in , quantified lexical and phonetic similarities as percentages of shared forms across sites, revealing correlations between linguistic and geographic distances that support diffusive wave patterns over isolation-by-distance. For instance, William Labov's 1972 study of analyzed shifts, like the raising of /æ/ before nasals, propagating as social waves from lower-middle to upper-middle classes, with higher socioeconomic groups leading the innovation while lower strata lagged, illustrating stratified diffusion akin to spatial waves. These tools have expanded into computational dialectometry, enabling precise mapping of wave gradients. Post-2000 advancements integrate geographic information systems (GIS) with the wave model to visualize contemporary continua dynamically, as in the project, which overlays digital layers of phonetic and lexical data to simulate innovation spreads across . European atlases, such as those from the Atlas der deutschen Mundarten, now use GIS to interpolate densities and predict wave trajectories based on and , facilitating real-time analysis of ongoing changes in urban-rural interfaces. This approach enhances the model's applicability to modern contact zones, revealing how accelerates wave propagation in multilingual regions.

In Reconstructing Language Relationships

The wave model facilitates the reconstruction of proto-languages by identifying clusters of shared innovations that spread across dialect continua, allowing linguists to posit "linkages"—networks of partially related s shaped by both descent and —rather than strictly hierarchical proto-languages derived from tree models. This approach is particularly useful for language families with extensive contact histories, where innovations do not align neatly with bifurcating branches. In of , Malcolm Ross applied this method to group over 20 potential lineages into linkages based on systems and shared morphological features, reconstructing historical networks that account for areal alongside genetic inheritance. The model also addresses borrowing and convergence by modeling substrate influences as waves of contact that propagate linguistic features across related varieties. For instance, in (a branch of Uralic), early Indo-European loans—such as terms for and —spread through successive contact zones in , reflecting convergence driven by population movements and trade rather than direct . This wave-like explains why certain phonological and lexical traits appear irregularly distributed, helping to distinguish borrowed elements from inherited ones in historical . A specific application appears in the Indo-European periphery, where Anatolian branches like Hittite and Luwian exhibit wave-like contacts with , evidenced by shared syntactic patterns in argument structure and clause organization that suggest prolonged areal influence rather than pure genetic descent. These interactions, occurring in ancient from the onward, demonstrate how the wave model captures hybrid developments, such as calques and structural alignments, that challenge traditional family-tree assumptions. As a methodological tool, the has been adapted for wave scenarios by tracing trajectories—the boundaries of linguistic features—backward through time to infer original paths and points of origin. This involves mapping shared retentions and innovations across continua to reconstruct the spatial dynamics of change, providing a more nuanced view of historical relationships than linear descent alone. In the , studies integrated phylogenetic s, such as Neighbor-Net algorithms, to visualize combined and signals in language expansions, notably in the Bantu family where analyses revealed reticulate patterns of diffusion during the 3,000-year spread across , highlighting zones of intense contact that overlaid genetic subgroups.

Legacy and Criticisms

Influence on Modern Linguistic Theories

In , the wave model's emphasis on gradual diffusion through social networks has profoundly influenced contemporary analyses of variation, particularly in urban settings. William Labov's seminal work in Principles of Linguistic Change, Volume 2: Social Factors (2001) adapts the wave metaphor to describe "waves of change" propagating through communities, where innovations originate in specific social strata—such as younger or lower-middle-class speakers—and spread outward via interpersonal interactions, incorporating factors like , , and neighborhood dynamics. This framework has enabled variationist studies to model real-time shifts in dialects, such as the Northern Cities Vowel Shift in , by treating changes as successive waves influenced by rather than abrupt bifurcations. The integration of in further extends the wave model by representing language diffusion as reticulated graphs rather than strict trees, capturing horizontal transfers akin to wave overlaps. For instance, Ringe, Warnow, and Taylor's 2002 analysis of using perfect phylogeny algorithms revealed significant reticulation—non-tree-like mixing—due to areal contacts, aligning with wave theory's prediction of innovations spreading across dialect continua without clear boundaries. Subsequent graph-based simulations in the have quantified this by modeling diffusion probabilities on social networks, demonstrating how waves can create linguistic areas through repeated contacts, as seen in simulations of Germanic language divergence. Post-2000 developments in have drawn on wave principles to conceptualize rhizomatic structures in and , where changes propagate non-hierarchically like interconnected roots. Pieter Muysken's 2008 edited volume From Linguistic Areas to Areal Linguistics applies this to contact scenarios, arguing that wave-like fosters shared features across unrelated languages in regions like the , emphasizing emergent patterns from ongoing interactions over static genealogies. This rhizomatic approach has informed studies of genesis, viewing pidgins as initial waves that evolve through layered contacts in colonial contexts. In , the wave model informs real-time tracking of slang propagation via , with 2020s projects leveraging data to visualize lexical innovations diffusing geographically and socially. For example, analyses of neologisms show patterns from urban youth networks to broader online communities, using tweet corpora to map spreads, akin to historical waves but accelerated by digital connectivity. Post-2014 research has incorporated simulations to model on , addressing documentation gaps by forecasting trajectories in varied settings. Recent advancements as of 2024 include models for predicting wave-like in globalized multilingual contexts, enhancing applications in computational .

Limitations and Debates

One key limitation of the wave model is its lack of predictive power regarding the speed or direction of linguistic innovations, as it relies on retrospective descriptions of diffusion patterns without incorporating sufficient or demographic data to forecast changes. This qualitative approach, while useful for mapping areal features, hinders precise quantification and modeling compared to the more structured, testable predictions offered by models in . The model has sparked ongoing debates, particularly its tension with Neogrammarian principles of regularity, which emphasize exceptionless genetic descent; waves accommodate diffusion-induced exceptions, but critics such as argued in the 1886 edition of his Prinzipien der Sprachgeschichte that this underemphasizes the primacy of vertical inheritance over horizontal borrowing. Modern criticisms highlight the wave model's overemphasis on , which can obscure deep-time genetic relationships captured better by trees; for instance, Gray and Atkinson's 2003 analysis of Indo-European core vocabulary using Bayesian demonstrated strong tree-like signals in basic , suggesting diffusion plays a secondary role in long-term divergence. Unresolved issues persist in distinguishing wave-based borrowing from signals of shared ancestry, complicating reconstructions without additional ; this has prompted calls for hybrid approaches integrating tree and wave elements, as seen in Bayesian phylogenetic networks developed in the that model both and lateral transfer. Post-2014 debates have also raised ethical concerns about applying and models to colonial language shifts, where analyses risk overlooking power imbalances and cultural in contact scenarios without community involvement or decolonial frameworks.

References

  1. [1]
    [PDF] Trees, Waves and Linkages: Models of Language Diversification
    Mar 19, 2015 · An important implication of the Wave Model is that a given language can perfectly well belong to several partially overlapping subgroups. A ...
  2. [2]
    (PDF) Wave Model - ResearchGate
    Nov 26, 2024 · Wave models of language change emerged during the latter half of the nineteenth century in Indo-European historical linguistics as an ...
  3. [3]
    August Schleicher's (1863) Indo-European Stammbaum, or ...
    The linguist August Schleicher (1821-1868) published a paper depicting an Indo-European language tree entitled Die Darwinshce Theorie und die ...
  4. [4]
    Die verwantschaftsverhältnisse der indogermanischen sprachen
    Jun 27, 2009 · Die verwantschaftsverhältnisse der indogermanischen sprachen ; Publication date: 1872 ; Topics: Indo-European languages ; Publisher: Weimar, H.Missing: wave | Show results with:wave
  5. [5]
    Trees, Waves and Linkages: Models of Language Diversification
    Schmidt, Johannes. 1872. Die Verwandtschaftsverhältnisse der indogermanischen Sprachen. Weimar: Hermann Böhlau. Schuchardt, Hugo. 1885. Über die Lautgesetze: ...
  6. [6]
    [PDF] Waves in computer simulations of linguistic diffusion - swanrad.ch
    The wave model or wave theory goes back to the work of Hugo Schuchardt ... then to analyze the spread of the High German consonant shift not as a wave which ...
  7. [7]
    Ueber die Lautgesetze. Gegen die Junggrammatiker. Berlin, 1885.
    Schuchardt, Hugo: Ueber die Lautgesetze. Gegen die Junggrammatiker. Berlin, 1885, S. 33. In: Deutsches Textarchiv <https://www.deutschestextarchiv.de/ ...
  8. [8]
    Schuchardt the iconoclast | Language or Dialect ... - Oxford Academic
    In 1885, he wrote a radical critique of the neogrammarians, reproaching their too systemic and mechanistic approach to language change. By this time, Schuchardt ...
  9. [9]
    History of the Institute - Research Center Deutscher Sprachatlas
    Georg Wenker sends a questionnaire with 42 short “folk” sentences to schools in the Rhineland, which he translated into the respective local dialects with the ...
  10. [10]
    [PDF] Three Waves of Variation Study: The Emergence of Meaning in the ...
    Jun 19, 2012 · The first wave of variation studies began with. Labov's (1966) study of the Social Stratification of English in New York City. Labov's main.
  11. [11]
  12. [12]
    Languages of the New Guinea Region (Chapter 28)
    Ross, Malcolm, 2001. Contact-induced change in Oceanic languages in north-west Melanesia. In Aikhenvald, Alexandra Y. and Dixon, R. M. W. (eds), Areal ...
  13. [13]
    [PDF] papers in honour of Malcolm Ross - ANU Open Research
    May 23, 2020 · Title: Discovering history through language : papers in honour of Malcolm Ross / editor Bethwyn Evans. ... Austronesian expansion. Roger ...
  14. [14]
    [PDF] A Comparative Grammar of the Early Germanic Languages - Loc
    May 20, 2016 · ... wave theory', first posited by J. Schmidt. 1872).3 An example of this ... High German Consonant Shift. (§6.21); and (f) elimination of ...
  15. [15]
    Urban Levantine Dialectal Features and the Levantine ... - jstor
    The Levantine-Mesopotamian dialect continuum is the result of important linguis- tic contacts through the centuries, and the existence of an Aramaic ...
  16. [16]
    [PDF] DELEUZE GUATTARI - Libcom.org
    3 Root, radicle, and rhizome—Issues concerning books—The One and the Multiple—Tree and rhizome—The geographical directions,. Orient, Occident, America—The ...
  17. [17]
    Aryan and Non-Aryan in India - Project MUSE
    Others argue that retroflexion in Indo-Aryan is a result of contacts with Dravidians ... sounds in terms of Dravidian influence on Indo-Aryan. This shift ...
  18. [18]
    [PDF] The Balkan Languages and Balkan Linguistics
    Europe as a linguistic area (Sprachbund)” (pp. v—vi). As Hamp (1977) ... Various linguists make much of the fact that the postposed definite article does not occur.
  19. [19]
    Linguistic Distances in Dialectometric Intensity Estimation
    Jul 4, 2014 · The so-called Levenshtein distance is a very popular phonetic distance measure. It is a string edit distance between two strings of phonetic ( ...
  20. [20]
    Advances in Dialectometry - Annual Reviews
    Jul 28, 2014 · Séguy was sadly lost to dialectometry through an accident, but his work was continued and deepened enormously by Hans Goebl, who may be regarded ...
  21. [21]
    [PDF] A QUANTITATIVE STUDY OF SOUND CHANGE IN PROGRESS
    in Paris (1960), Labov in Martha's Vineyard (1963) and. New York City (1966); more recently on-going change has been located by Cook in Salt Lake City (1969) ...
  22. [22]
    (PDF) Advances in Dialectometry - ResearchGate
    Dialectometry applies computational and statistical analyses within dialectology, making work more easily replicable and understandable.
  23. [23]
    Geography and Spatial Analysis in Historical Linguistics
    Aug 26, 2014 · Traditionally, the geographic component of dialect geography involved plotting data from questionnaires on the map pages of linguistic atlases ...
  24. [24]
    (PDF) Linguistic atlases. Traditional and modern - ResearchGate
    Forschungszentrum Deutscher Sprachatlas. Wenker, Georg. 1881 Sprach-Atlas ... Strasbourg: Trübner. Wenker, Georg. 1889⫺1923 Sprachatlas des deutschen Reichs ...
  25. [25]
    [PDF] Spatio-temporal modelling of language change
    The wave model describes the spread of linguistic features in space between languages that are in contact. More recently, both concepts. – trees and waves ...
  26. [26]
    (PDF) Historical language contact between Indo-European and ...
    Aug 6, 2025 · South and East IE languages have striking correspondences with Semitic, which is also poor in oblique experiencers and in impersonal ...
  27. [27]
    Principles of Linguistic Change, Volume 2: Social Factors | Wiley
    This volume presents the long-anticipated results of several decades of inquiry into the social origins and social motivation of linguistic change.Missing: waves | Show results with:waves
  28. [28]
    Principles of Linguistic Change, Volume II: Social Factors
    Research by Labov (2001) illustrates how dialects often correlate with social stratification, affecting individuals' opportunities and social mobility.Missing: waves | Show results with:waves
  29. [29]
    [PDF] indo-european and computational cladistics1 - Rice University
    This paper uses a new computational method to recover the first-order subgrouping of the Indo-European family, based on a 'perfect phylogeny' algorithm.Missing: wave | Show results with:wave
  30. [30]
    Splits or waves? Trees or webs? How divergence measures and ...
    Dec 12, 2010 · The 'network' type of phylogenetic method offers this, so we review recent applications to language data. Most have used lexical data, encoded ...
  31. [31]
    Social Networks of Lexical Innovation. Investigating the ... - Frontiers
    In this paper, I conduct a longitudinal study of the spread of 99 English neologisms on Twitter to study their degrees and pathways of diffusion.Missing: wave projects
  32. [32]
    Diffusion of Lexical Change in Social Media | PLOS One
    Nov 19, 2014 · Language change has long been an active area of research, and a variety of theoretical models have been proposed. In the wave model, linguistic ...
  33. [33]
    Using hybridization networks to retrace the evolution of Indo ...
    Sep 6, 2016 · Most of the IE language studies use the traditional phylogenetic tree model to represent the evolution of natural languages, thus not taking ...
  34. [34]
    Ethics in Linguistics - Annual Reviews
    In linguistics, ethics has long encompassed matters typically covered under regulatory oversight, but it is increasingly understood as relational and ...