Human settlement into new regions is typically accompanied by waves of animal extinctions, yet we have limited understanding of how human communities perceived and responded to such ecological crises. The first megafaunal extinctions in New Zealand began just 700 years ago, in contrast to the deep time of continental extinctions. Consequently, indigenous Māori oral tradition includes ancestral sayings that explicitly refer to extinct species. Our linguistic analysis of these sayings shows a strong bias towards critical food species such as moa, and emphasizes that Māori closely observed the fauna and environment. Temporal changes in form and content demonstrate that Māori recognized the loss of important animal resources, and that this loss reverberated culturally centuries later. The data provide evidence that extinction of keystone fauna was important for shaping ecological and social thought in Māori society, and suggest a similar role in other early societies that lived through megafaunal extinction events. . . . . . . . Keywords Cultural evolution Indigenous resource management Megafauna Moa Oceania New Zealand Maori Socio-ecological systems Traditional ecological knowledge Introduction Island ecosystems act as natural laboratories for understanding Electronic supplementary material The online version of this article the processes of extinction (Carlquist 1974;Prebble and (https://doi.org/10.1007/s10745-018-0004-0) contains supplementary material, which is available to authorized users. Wilmshurst 2009), including the role of, and impact upon, human communities. The Pacific Ocean was one of the last * Priscilla M. Wehi regions settled by humans, and its extinction timeline is there- email@example.com fore more recent than on the continents. However, even in the Pacific, most major extinction phases occurred well before Murray P. Cox written records. Consequently, although we know a great deal M.P.Cox@massey.ac.nz about the science of extinction events from archaeological and Tom Roa other data, we still understand little about how human com- firstname.lastname@example.org munities perceived, and responded to, the resulting ecological Hēmi Whaanga crises, or the development of community conservation ‘rules’ email@example.com and actions by these communities. By applying a combination of quantitative and qualitative linguistic methods, we have Centre for Sustainability, University of Otago, PO Box 56, explored indigenous responses to major faunal extinctions in Dunedin 9054, New Zealand one commonly used form of oral tradition, whakataukī.In Te Pūnaha Matatini, Manaaki Whenua – Landcare Research, 764 New Zealand, settled as recently as AD 1280 (Wilmshurst et Cumberland St, Private Bag 1930, Dunedin 9054, New Zealand al. 2008; Perry et al. 2014), the strong oral tradition among Te Pūnaha Matatini, Statistics and Bioinformatics Group, Institute of indigenous Māori includes a series of ancestral sayings Fundamental Sciences, Massey University, Private Bag 11 222, (whakataukī) that provide glimpses into the island’s early ex- Palmerston North 4442, New Zealand 4 tinction events and their importance as cultural signposts. In Faculty of Māori and Indigenous Studies, University of Waikato, generalized form, our findings may shed light on how other Private Bag 3105, Hamilton 3240, New Zealand Hum Ecol societies became aware of, and responded to, earlier animal and content of whakataukī over time would reflect rapid and extinctions across Eurasia and the Americas. ongoing socioenvironmental evolution associated with new In common with other Pacific island environments, the settlement (Wray and Grace 2007). We expected that arrival of humans in New Zealand resulted in high rates of whakataukī length would reflect social change. Specifically, extinction among its predominantly avian fauna (Duncan et we hypothesized that early settlement in close-knit family al. 2002;Bromham et al. 2012). The species most affected groupings, where relevance and implied contexts were well- were those especially vulnerable to human hunters, including understood within the group would result in shorter the group of large, flightless birds known as moa [Aves: whakataukī; whereas in the changing environments of rapid Dinornithiformes] (Holdaway 1989; Worthy 1997; Cassey settlement expansion and the formation of larger sub-tribal 2001). Detailed reconstruction of Holocene bird fauna indicates and tribal groupings, alliances, and warfare the context of that approximately 28 land bird species became extinct on the whakataukī would be less well understood and they would, two main islands of New Zealand in the 500 years between on average, increase in length (Wray and Grace 2007). Finally, initial settlement by Māori (c. AD 1280) and first European we predicted that major negative environmental change, such contact (AD 1769) (Worthy and Holdaway 2002; Tennyson as the loss of critical megafaunal food species, should drive a 2006;Wood 2013). Berkes (2008) argues that ecological crises progression from immediate observations of loss to a general- trigger learning in human communities, which in turn shapes ized understanding of the causes of extinction and finally to subsequent resource management practices. However, few stud- the deployment of explicit ecological management practices ies have explored the development of conservation learning in in keeping with the development of conservation practice response to ecological crises, and despite the rapid loss of major (Best 1904;Berkes 2008; Bowman et al. 2015). avian megafauna in New Zealand less than 700 years ago (Holdaway and Jacomb 2000; Tennyson 2006;Allentoft et al. 2014;Perry et al. 2014) how this ecological change affected Materials and Methods cultural learning remains essentially unknown. Quantitative analysis of linguistic markers to determine the Prior to European arrival in the late seventeenth century, timing and evolution of manuscripts has been increasingly Māori was a pre-literate language with a strongly developed employed over the last 20 years (Barbrook et al. 1998; oral tradition and a large unwritten literature of songs, poetry, Spencer et al. 2004; Eagleton and Spencer 2006;Howeand and proverbs as in many indigenous cultures. This tradition Windram 2011). More recently, these analyses have been ex- uses whakataukī, as well asnarratives(pūrākau)that contain tended to investigate the cultural legacies of folk tale records, philosophical thought, metaphor (kupu whakarite), epistemo- some of which likely originated before the emergence of written logical constructs, cultural codes, worldviews, and song records (da Silva and Tehrani 2016). The contexts in which (waiata) in everyday life. Whakataukī formed an important communications take place also shape linguistic form and length part of this tradition (Mead and Grove 2001; Wehi 2009). It (Wray and Grace 2007). Linguistic theory suggests that com- is important to note that although the translation of munication within close family groupings is dominated by im- whakataukī as ‘ancestral sayings’ suggests an association with plicit meaning. This contrasts with the cues that dominate exo- historic oral tradition, these sayings are still widely used by teric language used by distantly related groups, such as increased orators and speakers of Māori today. length and transparency (Wray and Grace 2007). Ecological information conveyed in ancestral sayings during early human Dataset settlement phases in a new land is thus likely embedded implic- itly in short phrases. Here, we use linguistic, historical and struc- European settlement in NZ was initiated shortly after 1800, tural cues in a body of whakataukī to analyze the development and gathered momentum in the second half of the nineteenth of socioecological thought over a period of c.650 years from the century (Fig. 2). During this period, many early European time of Polynesian arrival in New Zealand. ethnographers recorded and compiled Māori oral tradition, We first hypothesised that large bodied animals, such as the including Grey, Colenso, Smith, White, Williams, Best, and flightless moa, would predominate in Māori whakataukī if Firth (Supplementary Materials). These source materials, food sources were an important preoccupation for these set- along with other archived records, were comprehensively tlers. Avian body size is a significant predictor of hunting compiled, revised, translated, and interpreted by Mead and intensity across the Pacific (Duncan et al. 2002), and moa – Grove (2001) with the later addition of translations and inter- a group of ratites that ranged from the size of a turkey to much pretations. We used this pariemological dataset of 2669 larger than an ostrich – were a primary food and tool resource whakataukī as our primary dataset, supplemented by similar for the Polynesian ancestors of the indigenous Māori people entries from other compilations (total n = 3421; see electronic of New Zealand (Anderson 1989), given the lack of native supplementary materials for details). From this dataset, we mammals in New Zealand. Second, we predicted that the form analysed 657 whakataukī that explicitly refer to fauna. Hum Ecol Time Attribution on state television in Māori) that aired in late November/ December 2015. The program’s content focuses on topics of Drawing from Mead’s(1984) classification of the temporal de- national significance to the targeted Māori audience. velopment of Māori artistry, and Davidson’s(1984) cultural The mobile unit of the National Broadcasting Service made phases, each faunal whakataukī was assigned to one of five broad recordings of Māori speakers between 1946 and 1948. The Unit time periods: before AD 1350 (pre-dating Māori settlement of made three tours: two in the North Island, covering the West New Zealand), 1350-1500 (the very early New Zealand settle- Coast from Wanganui to Waitara in late 1946, and the Waikato ment period), 1500-1650 (rapid settlement expansion), 1650- and Thames Valley districts in 1947, and one in the South Island 1800 (intertribal fighting and early European contact) and post- to the Otago region in 1948. Sentences were selected in the 1800(theperiodfollowingEuropeancontact).Thesetimeperiods following way. First, for the speakers born in the 1870s, we equate approximately, in English literature terms, to the early demarcated each new topic of conversation, and selected the Anglo-Saxon period of Beowulf (~1000 AD); to that of first sentence after the third sentence that was between 15 and Chaucer (~1343-1400); Shakespeare (1564-1616); Jonathan 50 words in length. Each sentence was checked for any obvious Swift (1667-1745); and Charles Dickens (1812-1870). To assign words or dates that would provide the informant with explicit faunal whakataukī to these temporal categories, we called upon temporal information (e.g., references to World War II). If the TR, a Māori linguist, translator, historian, and native speaker, to first example was deemed unsuitable, or if parts were inaudible, derive estimated chronological dates for the whakataukī without the next example of 15-50 words was selected and checked for anypriorknowledge ofthe hypothesesof the study (see belowfor suitability. For the second set of speakers, we excluded validation) so that we could establish relationships both between sentences spoken by the interviewer as these are normally different versions of the same whakataukī,andbetween scripted rather than representing impromptu speech. We then whakataukī with different content (Table 1). selected the first sentence of the interviewee of between 15 TR used linguistic and structural cues, vocabulary identifi- and 50 words in length. A maximum of five sentences was cations, historical contexts and embedded references to ances- selected from a single speaker (Table S4). tor names, events, and genealogies to make these temporal Language expert TR achieved a hit rate of 81% and error rate assignments (Roa 2016)(Table 1). For example, specific his- of 19%, yielding a sensitivity index d’ of 1.76. Sentences from torical details form the context for the whakataukī on fish the early period were assigned with slightly more accuracy shown in Table 1. In other whakataukī, estimated dates of (86%) than sentences from the late period (76%). The null species arrivals (e.g., for pītongatonga, after European arrival) hypothesis that the language expert assigns sentences randomly and old and modern word usages and transliterations (e.g., pī to the early and late periods can therefore be rejected both for and heihei) were used to cross reference dating. the set of early sentences (χ = 13.4, P = 0.00025) and the set of late sentences (χ = 28.2, P = 0.00000011). Given this discrim- Validation of Time Period Assignments inatory power for two time periods separated by only 100 years, the whakataukī were likely assigned to five much wider time Scoping tests demonstrate that expert linguistic training is periods with similar or better accuracy. necessary for accurate dating over and above native-speaker and community elder status (Chipere 2000). As a result, we devised a blind validation test to confirm the ability of our Knowledge Development language expert and co-author (TR) that used Māori sentences with no temporal context taken from two time periods (late We classified whakataukī according to Berkes’ (2008)mech- nineteenth century, early twenty-first century) to obtain a reli- anisms of knowledge development (i.e., observation and mon- ability value for distinguishing early and late sentence struc- itoring, trial and error experimentation, learning from other tures dating from approximately a century apart. We presented places and times, and knowledge encoded in language and TR with a randomized set of sentences of comparable length other narratives) to examine key developments in traditional (approximately 15 – 50 words) from oral recordings in the knowledge, practice, and resource management over these Māori language, with 100 examples selected from each of time periods. In addition, we examined whakataukī using a two time periods (Table S4). model proposed by Crombie (1985a, b, 1987) and Whaanga Sentences were selected from six speakers of Māori who (2006) that identifies mechanisms of human thinking and lan- were recorded in the 1940s, but were born around the 1870s. guage structure based on cognitive processes. The model uses These are regarded as the landmark early Māori oral recordings four main distinctions (temporal, additive, associative, and and have been used extensively in other comparative studies causal) to explore connections between the development of (Keegan et al. 2014). These were compared with sentences knowledge through time and language structure shifts and takenfrom58Māori speakers who were recorded on Te time periods. To make these classifications, the model iden- Karere (a news and current affairs television program broadcast tifies discourse relations, and the presence and absence of Hum Ecol Table 1 Examples showing how linguistic cues, historical and cultural context, and identification of events and ancestor names inform chronological dating of whakataukī Example Theme English translation Time period Explanation (a) Moa 1. Mate ā-moa Extinction Dead as the moa 1500-1650 The first seven examples have similar content or core meanings, 2. Ka ngaro ā-moa Extinction This tribe will disappear 1800- but display shifts in language structure (to a greater or lesser te iwi nei like the moa extent). For instance, example (1) is the most compact and 3. Ka ngaro ā-moa Extinction The Māoriwillbecome 1800- the only example to use the stative verb mate (be dead) with te tangata extinct like the moa the ‘ā-prefix/noun’ combination to derive a stative adjective; 4. Ka ngaro i te Extinction Lost like the loss of the 1800- we believe this is the oldest of these eight whakataukī. ngaro o te moa moa Examples (2) and (3) are similar to example (1), using the 5. Kua ngaro i te Extinction Perished as the moa 1500-1650 ‘stative/ā-prefix/noun’ combination, but additional ngaro o te moa perished information is included: the stative verb ngaro (be missing, 6. Huna i te huna Extinction Hidden as the moa hid 1800- lost, hidden, extinct), the particle ka to mark tense/aspect ate moa and a nominal phrase to indicate subject – the difference in 7. Ko te huna Extinction It is like the disappearance 1650-1800 subject being te iwi nei ‘this tribe’ and te tangata ‘the i te moa! of the moa person/the Māori people’. Example (3) and the following 8. He puku moa! Ecological A stomach of a moa! 1350-1500 example express the widely held belief in the late nineteenth observation century that Māori would similarly become extinct. Examples 9. He rātā te rākau Ecological Arātā was the tree 1500-1650 (4) and (5) are structurally different from the first three itakahia ete moa observation trampled by the moa examples, but very similar to each other. They include the 10. He koromiko te Food Koromiko is the wood 1500-1650 stative verb ngaro and the prepositional phrase ite ngaro o wahie i taona ai preparation with which the moa te moa, but are marked with different tense/aspect particles te moa was cooked – ka vs. kua ‘perfect tense’. The perfect tense marker in this case implies the achievement of a state as the result of an event. On the other hand ‘Ka’ indicates the past. Thus, example (5) was used for people suddenly killed or carried off by death. Example (6), however, is not marked by any tense/aspect marker, it introduces the stative verb huna (be concealed, unnoticed) and includes the prepositional phrase i te huna a te moa. Mead suggests the saying expresses contempt at the poor concealment ability of moa. However, it also refers to the disappearance of a social group, or aspect of culture. This could alternatively suggest an element of disbelief that something so ubiquitous could disappear. Example (7) is structurally different from the rest: it uses the stative huna in an emphatic construction with the particle ko with the prepositional phrase ite moa. Example (8) describes an ecological observation in which the moa, a giant herbivore browser with up to 5 kg of gizzard stones that were used to break down fibrous plant material, were compared with gluttonous people, who similarly browse great quantities of food. Example (9) describes the rātā, a parasitic forest tree, that fails to stand upright independently, and is susceptible to trampling while young. For example (10), the boughs, leaves and flowers of the koromiko tree may have been used to cover the moa flesh when cooked in an umu (a ground oven). (b) Fish Hā!Heika poto Historical What! A short fish, 1500-1650 Specific historical details form the context for this saying, and can te ika nei! this one’ thus be used to help identify its chronology. Awakanoi of Ngāti Awa was slain by near Rūātoki. According to Best (1925), when the body was turned over to reveal its identity, the victor uttered this saying. It apparently meant he had hoped for a more prominent victim. Another explanation is that Ipuhue was disappointed that the victim had not provided more of a contest. (c) Chicken Ai pī Breeding Chicken breeding. before 1350 Referring to a prolific parent with numerous children, this whakataukī has been identified as pre-dating Māori settlement. Overpopulation is a recurring rationale for Māori departure from Hawaiki in oral tradition. Domestic chickens either were not carried, or did not survive on canoes during the journey to Aotearoa New Zealand. Hum Ecol relational signaling and encoding (i.e., coherence, cohesive weights, and all scientific names for birds that became extinct on devices, co-ordination, subordination, conjuncts, and lexis) the two main islands of New Zealand, have been provided in with linguistic and world knowledge (i.e., prior understand- Tables S1 and S2. ings) (Roa 2016; Supplementary Materials). Using these clas- We note that Māori names do not always map cleanly to sification techniques, we were able to establish relationships modern taxonomic units. Bird names in Fig. 2 may therefore between different versions of the same whakataukī,and refer to closely related and morphologically similar species, whakataukī with different content. such as moa, albatross, shag, kiwi, and kākāriki (native parakeets). Morphologically similar birds from the same ge- nus that are now described as different species on the North Word Frequencies and South Islands are often described by a single common name, such as ‘snipe,’‘saddleback,’ and ‘kōkako.’ In Fig. 2, Qualitative analyses included inspection of word frequencies some bird names have also been shortened for visual clarity: in whakataukī. All function words (such as ‘a,’‘the,’ and ‘in’) ‘bittern’ refers to the New Zealand bittern, ‘godwit’ to the bar- were removed from the dataset using standard UNIX opera- tailed godwit, ‘oystercatcher’ to the variable oystercatcher, tions (particularly the command line program ‘grep’). Word ‘heron’ to the white heron, and ‘quail’ to the extinct New frequencies were determined using an online word counting Zealand quail. Macrons are not shown for names on the plot. tool (http://www.textfixer.com/tools/online-word-counter. php; accessed April 2017). Historical Reports of Moa Bird Data We considered whether the whakataukī referring to moa and We used Dunning (2007) to determine mean weights for all their extinction that were identified as dating to the nineteenth bird species (Table S1, Supplementary Materials), and then century were in part a product of the intense scientific interest regressed these weights against total word occurrence for that that arose near that time about this bird. Anderson (1989)has species or group in the whakataukī faunal dataset. Prevalence inferred that the controversy generated within the scientific of bird species across New Zealand archaeological sites (Fig. community about the discovery of moa, and its apex predator 2) was obtained from Worthy (1999). the Haast’s Eagle, likely preceded or stimulated the recording of sayings about moa during the post European period, and Use of Māori and English Bird Names suggests that the moa were as common a symbol in public imagery then as the kiwi is today. However, a closer evalua- We have used bird and plant names that are common usage in tion of the evidence suggests otherwise (Supplementary New Zealand to refer to species reported in this paper. In some Materials). Governor George Grey asserted that Māori all cases these are Māori (e.g., kiwi, miro), and in some cases knew the word moa as ‘a bird well known to their ancestors.’ English (e.g., blue duck). We have not italicized these names, Grey also recorded the lament of Ikaherengutu that includes but provide the appropriate scientific name on first usage to references to moa, and was sung by Te Wherowhero on the assist the reader with identification. In addition, all scientific death of his brother (see Supplementary Materials for further names for birds represented in Fig. 1 along with their average details). Because Anderson’s survey of oral tradition was Fig. 1 Relative frequency of words in the faunal subset of Māori whakataukī,translatedhere into English. Function words have been removed Hum Ecol Fig. 2 Large birds (pictures 50 moa scaled according to size; Table S1) are discussed more often than small birds in whakataukī and are found at a larger number of ar- kaka chaeological sites. Moa are heavily represented in kereru whakataukī; a moa head only is weka shown due to their disproportion- shag kakariki ately large size. Birds represented heron tui in blue (i.e.moa and pouakai) be- takahe bellbird kokako came extinct prior to European harrier kiwi duck arrival – but other extinct birds do falcon saddleback not occur in the whakataukī and 5 snipe oystercatcher are thus not shown in the figure grey warbler fantail morepork (Table S2). Data from archaeo- albatross longtailed silvereye nz quail cuckoo logical sites are from Worthy (1999), shown with permission prion kingﬁsher pukeko fernbird shining cuckoo gull godwit huia gannet piopio robin 1 kakapo whitehead kea pouakai laughing owl 1 2 5 10 20 50 100 200 Number of Archaeological Sites primarily limited to narratives (pūrākau) a number of critical and 10.4 words for periods 4 and 5 respectively, t = −0.859, references to moa in waiata (song) and whakataukī were not P = 0.39). In addition, qualitative examination of early identified. whakataukī show they are dominated by implicit meanings indicative of closely related family groupings in contrast to Statistical Analyses whakataukī from later periods in which larger tribal group- ings, alliances, and warfare prevail. For example, the brief All statistical approaches were implemented in R (R phrase ‘ai pī’ references overpopulation as a recurring ratio- Development Core Team 2010). nale for Polynesian voyaging through the Pacific (Table 1). Moa form an archetypal group that are strongly over- represented in whakataukī as in the archaeological record Results (Figs. 2 and 3). Whakataukī about moa comprise 4.6% of all faunal whakataukī, and 9.8% of all whakataukī that specifically Among the faunal set of whakataukī, the most frequent word is mention birds. In striking contrast, other large extinct megafau- ‘food’(kai), indicating that subsistence was indeed a high priority na, such as the endemic geese and adzebills with adult weights (Fig. 1). As predicted, large birds are more frequently mentioned >15 kg, are not represented in the whakataukī.Indeed,their in whakataukī than small birds (n =657 whakataukī; r = 0.60, P original Māori names are largely lost (Fig. 4). Māori names for < 0.0001, Fig. 2), consistent with the same trend whereby large species that became extinct prior to European arrival are now birdsaremorecommonlyfoundinarchaeologicalsitesthansmall unknown, with the exception of words for moa and pouākai. birds (n =112 sites; r = 0.41, P = 0.0052; Fig. 2). This loss contrasts with the retention of names for avifauna that Few whakataukī were dated to pre-1350 or post-1800. became extinct after European arrival. Only one other extinct Nonetheless, whakataukī from the early settlement period bird with remarkable body size is mentioned – the Haast’sEagle are shorter (mean length 6.75 words) than whakataukī from (Aquila moorei), a giant raptor with a 3 m wingspan called the subsequent phase of rapid settlement expansion (AD pouakai or hokioi by Māori. This eagle is the only known apex 1500-1650, mean length 9.13 words; t = −5.10, P < predator of moa other than humans (Anderson 1989) and quick- 0.001). No increase in word length was seen when we com- ly followed the moa into extinction. pared period 4 (1650-1800, prior to European arrival) and Qualitative analysis of the dataset indicates that the nature period 5 (1800-, after European arrival) (mean lengths 9.56 and frequency of moa whakataukī are disproportionally Number of Whakataukī Hum Ecol Fig. 3 Whakataukī records in A All Records relation to time. a Whakataukī that refer to all fauna occur most frequently between AD 1500- 1650. b The abundance and likely importance of moa whakataukī varies through time, as shown by their relative proportion to all pre−1350 1350−1500 1500−1650 1650−1800 1800− faunal whakataukī during each time period. c The Polynesian founder population c. AD 1280 is generally estimated as <400 in B Moa Records size (Whyte et al. 2005), and population expansion was slow in comparison to rapid European 0.16 Ecology settlement in the nineteenth cen- 0.12 tury (Holdaway et al. 2014) 0.08 0.04 0.00 pre−1350 1350−1500 1500−1650 1650−1800 1800− Māori European 1280 1480 1680 1880 Calendar Years AD skewed toward two time periods (1350-1500, and post-1800) extinction (Fig. 3, Table 1). Ecological observations are espe- (Fig. 3). Three main themes emerge from the moa whakataukī: cially evident during the early settlement period. The two ecological information, food preparation, and concerns about peaks of whakataukī about moa reflect i) moa ecology and Avian extinctions pre−1350 1350−1500 1500−1650 1650−1800 1800− Time period Fig. 4 Most avian extinctions on mainland New Zealand occurred either for species in the 1350-1500 time period. See Table S2 for estimated ex- prior to AD 1500 (within 150 years of Māori arrival), or after European tinction dates. Key avian extinction periods occurred shortly after Māori arrival post-1800 AD. ‘Moa’ is treated here as a name for all nine species of and European settlement periods. Grey represents names birds for which moa, reflecting its indigenous usage; if this group of nine species is treated the Māori name is no longer known, and black represents bird species or as one taxonomic unit, the loss of names would be commensurately higher groups for which the Māori name has been retained Number of records Popula on (in thousands) Propor Number of Records and expansion First pā (palisaded fort) diseases and intertribal wars Hum Ecol their likely functional extinction around AD 1400-1450, and to be contemporaneous in the oral record. Diverse and emo- ii) archetypal links to ideas of extinction for Māori themselves tive language emphasises the ‘loss’ and ‘death’ of moa, sug- (Table 1) during the social upheaval that followed European gesting that the extinction of moa was widely noted and colonisation in the early to mid-nineteenth century. Almost all discussed. A later set of moa whakataukī appears after moa whakataukī post-1800 link moa disappearance to European arrival in the nineteenth century and almost uni- impending Māori extinction. In general, the whakataukī formly employs the loss of moa as a metaphor for the feared dataset demonstrates a trend from simple observation in the extinction of Māori. This re-mapping of whakataukī early settlement period to awareness of causal agency in later concerning the fifteenth century loss of moa to a much later periods (AD 1500-1800; Fig. 5), as expected for the develop- nineteenth century social crisis – the imminent and very real ment of traditional knowledge, resource management prac- threat of Māori biological and cultural extinction – powerfully tices, and conservation ‘rules’ (Berkes 2008). emphasizes the impact of moa on the cultural psyche of Māori. The frequency and content of these later whakataukī support the view that Māori were not only aware of the dismal end met by moa, but also that moa extinction came to serve as Discussion an archetypal exemplar for extinction more generally. Whakataukī word lengths in the early period support the The strong relationship between bird size, frequency in the predictions of esoteric and exoteric language theory (Wray oral record, and frequency in recorded material from archae- and Grace 2007), which expected linguistic forms and struc- ological sites indicates that despite the limitations of each of tures to change with communication context. Early these datasets Māori both targeted and talked about animal whakataukī indeed carry strongly embedded meanings that species (in this case, birds) that were important food resources. reflect close family groupings in contrast to whakataukī from Species that hold strong cultural significance, such as the later periods, with larger tribal groupings, which have greater white heron (Ardea modesta), are also well represented in transparency of meaning. That is, interactions with strangers, the whakataukī. While it is difficult to fully test our hypothesis language contact, and stratification of society all influence because, for example, small bones are less likely to be found language evolution and the way that ecological material is in archaeological deposits, the relationship shown here is clear presented. These findings align with the results of the dating for all larger bird species. methodology, providing confidence in the temporal assign- The frequency of references to moa in the whakataukī pro- ment of language cues and structures identified for each vides a powerful contrast to that of other large bird species that period. became extinct before European arrival. Some large species, Tracking changes in resource management practices such as adzebills, may have had restricted ranges, and there- proved challenging using whakataukī. Qualitative analysis re- fore might be expected to appear less. Nonetheless, the gen- veals a nuanced understanding of ecological relationships in eral absence of references to extinct birds in whakataukī em- faunal whakataukī from later time periods, for example noting phasises the language loss that frequently accompanies biodi- seasonal linkages between species (Table S3, Supplementary versity extinction (Maffi 2005). Materials). This is consistent with the finding that terms asso- Given the likely speed with which moa became extinct ciated with complex social relationships and structures, such (<150 years) (Perry et al. 2014) it is unsurprising that as chieftainship and territoriality, increase in frequency whakataukī on moa food preparation and on extinction appear Fig. 5 Whakataukī shift from a predominance of observations early on, towards a rise in causal sayings from 1500 to 1800, before a further small rise in associative observations after 1800 Hum Ecol through time (Wehi et al. 2013). Nevertheless, the progression Compliance with Ethical Standards from ecological observations to adaptive management is not The research was funded by the Royal Society of New Zealand from clearly reflected in this dataset. One possibility is that the Marsden funding to H.W and P.M.W. (12-UOW-093) and Rutherford nuances of adaptive management practices are too complex Discovery Fellowships to M.P.C. (RDF-10-MAU-001) and P.M.W. for the brief format of whakataukī (median 8 words, range 2- (RDF-14-LCR-001). 36 words, Tables 1 and S3). Notably, specific management practices are not explicitly discussed in any of the faunal Conflict of Interest The authors declare they have no conflict of interest. whakataukī even though environmental stewardship was Statement of Authorship PMW conceived of the study, and PMW, widely noted during the nineteenth century by early MPC and HW designed the study. PMW collected data on whakataukī European writers (Best 1904;Kawharu 2000;Kitsonand and extinct and living species, MPC carried out statistical analyses, TR Moller 2008). For instance, harvests of sooty shearwater carried out dating of whakataukī and modern texts and developed (tītī) have operated continuously for several centuries methods for language dating, and HW collected data on whakataukī and classified language structure. PMW wrote the first draft of the man- (Hawke et al. 2003) with defined sustainability practices uscript. All authors participated in data interpretation and contributed passed down inter-generationally (Kitson and Moller 2008) substantially to manuscript development and revisions. and the management of harvests falling largely under the re- Open Access This article is distributed under the terms of the Creative sponsibility of chiefs (Best 1904;Kawharu 2000). Our focus Commons Attribution 4.0 International License (http:// on faunal whakataukī may have excluded whakataukī that creativecommons.org/licenses/by/4.0/), which permits unrestricted use, mention ecosystem-wide management. Ecological manage- distribution, and reproduction in any medium, provided you give ment events and environmental frameworks may also be more appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. conspicuous in other forms of oral culture, such as storytelling (Bowman et al. 2015). Alternatively, observation may not necessarily have led to immediate or visible action, perhaps as reflected in the unexpected absence of most large extinct References birds in the whakataukī. Oral tradition, such as these whakataukī passed down by Allentoft, M. E., Heller, R., Oskam, C. L., Lorenzen, E. D., Hale, M. L., Māori, provide our only real glimpses into the ecological re- Gilbert, M. T. P., Jacomb, C., Holdaway, R. N., and Bunce, M. lationships and concerns of early settler populations, and pro- (2014). Extinct New Zealand Megafauna Were Not in Decline vide early human context to an otherwise relatively dry scien- Before Human Colonization. Proceedings of the National Academy of Sciences 111(13): 4922–4927. https://doi.org/10. tific record of extinction events. The whakataukī emphasise 1073/pnas.1314972111. that indigenous peoples are not simply passive actors against Anderson, A. S. (1989). Prodigious Birds - Moas and Moa-Hunting in an environmental backdrop but rather interact with the envi- Prehistoric New Zealand, Cambridge University Press, Cambridge. ronment in myriad ways that affect not only the species as- Barbrook, A. C., Howe, C. J., Blake, N., and Robinson, P. (1998). The semblages present but also the development of cultural values, Phylogeny of the Canterbury Tales. Nature 394(6696): 839–839. https://doi.org/10.1038/29667. ideas, and practices. As such, these whakataukī provide evi- Berkes, F. (2008). Sacred Ecology, Routledge, New York. dence of the links between cultural and biological diversity Best, E. (1904). Notes on the Custom of Rahui. Its Application and (Maffi 2005). Similar linguistic analysis of other indigenous Manipulation, as Also its Supposed Powers, its Rites, Invocations oral traditions globally could illuminate the development of and Superstitions. The Journal of the Polynesian Society 13(2): 83– socioecological world views and conservation learning in oth- er cultures, at least where extinction events are relatively re- Bowman, D. M. J. S., Gibson, J., and Kondo, T. (2015). Outback Palms: Aboriginal Myth Meets DNA Analysis. Nature 520(7545): 33–33. cent. They provide our closest available proxy to the thoughts https://doi.org/10.1038/520033a). and responses of the human communities that lived through Bromham, L., Lanfear, R., Cassey, P., Gibb, G., and Cardillo, M. (2012). the big megafaunal extinctions, which occurred very early in Reconstructing Past Species Assemblages Reveals the Changing human history on the continents. Patterns and Drivers of Extinction Through Time. Proceedings of the Royal Society B: Biological Sciences 279(1744): 4024–4032. https://doi.org/10.1098/rspb.2012.1437. Carlquist, S. J. (1974). Island Biology, Columbia University Press, New Acknowledgements Trevor Worthy (Flinders University) provided addi- York. tional information on bird species from archaeological sites throughout Cassey, P. (2001). Determining Variation in the Success of New Zealand New Zealand. We thank Kevin Laland (University of St Andrews) for Land Birds. Global Ecology and Biogeography 10(2): 161–172. stimulating discussion, Alan Tennyson (Te Papa National Museum, New https://doi.org/10.1046/j.1466-822x.2001.00224.x. Zealand) for advice on extinct bird abundance in prehistoric New Zealand, and Pou Temara (University of Waikato) for assisting with tests Chipere, N. (2000). Processing embedding in humans and connectionist of linguistic and historical cues. David Barrell, Emma Brooks, Andreea models [Doctor of Philosophy (PhD)]. Cambridge, England, Calude, Paul Scofield and Dan Tompkins provided constructive com- University of Cambridge. ments on an earlier draft of the manuscript. David Wheadon improved Crombie, W. (1985a). Discourse and Language Learning: A Relational the illustrations in Figs. 2 and 3. Approach to Syllabus Design, Oxford University Press, Oxford. Hum Ecol Crombie, W. (1985b). Process and Relational Discourse and Language Perry, G. L. W., Wheeler, A. B., Wood, J. R., and Wilmshurst, J. M. Learning, Oxford University Press, Oxford. (2014). A High-Precision Chronology for the Rapid Extinction of Crombie, W. (1987). Free Verse and Prose Style, Croom Helm, London. New Zealand Moa (Aves, Dinornithiformes). Quaternary Science da Silva, S. G., Tehrani, J. J. (2016). Comparative Phylogenetic Analyses Reviews 105: 126–135. https://doi.org/10.1016/j.quascirev.2014. Uncover the Ancient Roots of Indo-European Folktales. Royal 09.025. Society Open Science 3(1). https://doi.org/10.1098/rsos.150645. Prebble, M., and Wilmshurst, J. M. (2009). Detecting the Initial Impact of Davidson, J. M. (1984). The Prehistory of New Zealand, Longman Paul, Humans and Introduced Species on Island Environments in Remote Auckland. Oceania Using Palaeoecology. Biological Invasions 11(7): 1529– Duncan, R. P., Blackburn, T. M., and Worthy, T. H. (2002). Prehistoric 1556. https://doi.org/10.1007/s10530-008-9405-0. Bird Extinctions and Human Hunting. Proceedings of the Royal R Development Core Team (2010). R: A Language and Environment for Society B 269: 517–521. https://doi.org/10.1098/rspb.2001.1918. Statistical Computing, R Foundation for Statistical Computing, Dunning, J. B. (2007). CRC Handbook of Avian Body Masses, CRC Vienna https://www.r-project.org/. Press, Hoboken. Roa, T. (2016) An exploration of the role of semantic relations in the Eagleton, C., and Spencer, M. (2006). Copying and Conflation in theory and practice of translation (with special reference to Geoffrey Chaucer’s Treatise on the Astrolabe: A Stemmatic English/Māori and Māori/English translation) [Doctor of Analysis Using Phylogenetic Software. Studies in History and Philosophy (PhD)]. Hamilton, New Zealand, University of Waikato. Philosophy of Science Part A 37(2): 237–268. https://doi.org/10. Spencer, M., Davidson, E. A., Barbrook, A. C., and Howe, C. J. (2004). 1016/j.shpsa.2005.08.020. Phylogenetics of Artificial Manuscripts. Journal of Theoretical Hawke, D., Newman, J., Moller, H., and Wixon, J. (2003). A Possible Biology 227(4): 503–511. https://doi.org/10.1016/j.jtbi.2003.11. Early Muttonbirder’s Fire on Poutama, a Rakiura Titi Island, New 022. Zealand. Journal of the Royal Society of New Zealand 33(2): 497– Tennyson, A. (2006). Extinct Birds of New Zealand, Te Papa Press, 507. Wellington. Holdaway, R. N. (1989). New Zealand’s Pre-human Avifauna and its Wehi, P. M. (2009). Indigenous Ancestral Sayings Contribute to Modern Vulnerability. New Zealand Journal of Ecology 12(Suppl): 11–25. Conservation Partnerships: Examples Using Phormium tenax. Holdaway, R. N., and Jacomb, C. (2000). Rapid Extinction of the Moas Ecological Applications 19(1): 267–275. https://doi.org/10.1890/ (Aves: Dinornithiformes): Model, Test, and Implications. Science 07-1693.1. 287: 2250–2254. https://doi.org/10.1126/science.287.5461.2250. Wehi,P.M.,Cox,M.,Roa,T.,andWhaanga,H.(2013).Marine Holdaway, R. N., Allentoft, M. E., Jacomb, C., Oskam, C. L., Beavan, N. Resources in Oral Tradition: He Kai Moana, He Kai Mā Te R., and Bunce, M. (2014). An Extremely Low-Density Human Hinengaro. Journal of Marine and Island Cultures 2: 59–68. Population Exterminated New Zealand Moa. Nature Whaanga, J. P. (2006). Case roles/ relations and discourse relations: A Communications 5(5436): 1–8. https://doi.org/10.1038/ Māori language-based perspective [Doctor of Philosophy (PhD)]. ncomms6436). Hamilton, New Zealand, University of Waikato. Howe, C. J., and Windram, H. F. (2011). Phylomemetics—Evolutionary Whyte, A. L. H., Marshall, S. J., and Chambers, G. K. (2005). Human Analysis Beyond the Gene. PLoS Biol 9(5): e1001069. https://doi. Evolution in Polynesia. Human Biology 77(2): 157–177. org/10.1371/journal.pbio.1001069. Wilmshurst, J. M., Anderson, A. J., Higham, T. F. G., and Worthy, T. H. Kawharu, M. (2000). Kaitiakitanga: A Maori Anthropological (2008). Dating the Late Prehistoric Dispersal of Polynesians to New Perspective of the Maori Socio-Environmental Ethic of Resource Zealand Using the Commensal Pacific Rat. Proceedings of the Management. The Journal of the Polynesian Society 109(4): 349– National Academy of Sciences of the United States of America 370. 105(22): 7676–7680. https://doi.org/10.1073/pnas.0801507105. Keegan, P. J., Watson, C. I., Maclagan, M., and King, J. M. (2014). Sound Wood, J. R. (2013). Extinction - New Zealand, 500 years ago. In change in Māori and the formation of the MAONZE project. In MacLeod, N., Archibald, J. D., and Levin, P. (eds.), Grzimek’s Onysko, A., Degani, M., and King, J. M. (eds.), He Hiringa, He Animal Life Encyclopedia, Chicago, Booklist Publications, pp. Pūmanawa: Studies on the Māori Language, Huia, Wellington, pp. 595–604. 33–54. Worthy, T. H. (1997). The Identification of Fossil Eudyptes and Kitson, J. C., and Moller, H. (2008). Looking After Your Ground: Megadyptes Bones at Marfells Beach, Marlborough, South Island. Resource Management Practice by Rakiura Maori Titi Harvesters. New Zealand Natural Sciences 23: 71–85. Papers and Proceedings of the Royal Society of Tasmania 142(1): Worthy, T. H. (1999). What’s on the Menu? New Zealand Journal of 161–176. Archaeology 199: 125–160. Maffi, L. (2005). Linguistic, Cultural, and Biological Diversity. Annual Worthy, T. H., and Holdaway, R. N. (2002). The Lost World of the Moa: Review of Anthropology 34: 599–617. https://doi.org/10.1146/ Prehistoric Life of New Zealand, Canterbury University Press, annurev.anthro.34.081804.120437. Christchurch. Mead,H.M.(1984).TeMāori – Māori Art from New Zealand Wray, A., and Grace, G. W. (2007). The Consequences of Talking to Collections, Heinemann, Auckland. Strangers: Evolutionary Corollaries of Socio-cultural Influences on Mead, S. M., and Grove, N. (2001). Ngā Pēpeha a Ngā Tīpuna: The Linguistic Form. Lingua 117(3): 543–578. https://doi.org/10.1016/j. Sayings of the Ancestors, Victoria University Press, Wellington. lingua.2005.05.005.
– Springer Journals
Published: Jun 4, 2018