Language Extinction Triggers the Loss of Unique Medicinal

Downloaded by guest on September 29, 2021 ot mrc,nrhetAaoi,adNwGie being in Guinea services New medicinal and all the Amazonia, in of northwest uniqueness 84% America, and linguistic 91%, North of 73%, pattern with strong regions, a exhibits plants a of (e.g., System). subcategory combination Digestive medicinal the a as and service” Methods ). species plant and plant “medicinal (Materials a languages defined indigenous We 236 asso- services with plant 12,495 ciated and north- data species these three plant (11), Together, medicinal analyzed (13). 3,597 America Guinea span we New this, North and (12), of for Amazonia west extent datasets the ethnobotanical assess large To shared. knowl- such is whom lan- with edge be indigenous languages still surviving would some other knowledge if in plant safeguarded even medicinal their knowledge is, extinct, That go languages, guages high. different be would speak resilience that (10). among groups widely orally shared indigenous is knowledge medicines about transmit resilience knowledge if cultures its Therefore, indigenous for Most implications important (9). has this services been medicinal how has extinction and ecological knowledge by result unique posed 1). may (Fig. that limited linguistically languages to indigenous of compares of risk of loss the loss discovery of the the poten- understanding the in which our full and however, to the people far, degree limit to So uses. may contributions unanticipated nature’s century of coming (7, plant tial the crops useful geographic in and within reductions the plants and species constrain extinction endemic language associated will Together, human-utilized On 8). strongly change many (6). of global is plants ranges about hand, disuse knowledge other language extinctions indigenous the species hand, in and one decreases (3) with On loss 5). language (4, increas- by in is encoded threatened however, knowledge, is ingly Indigenous (1)—that (2). benefits languages health their significant confers that I languages indigenous than knowledge be medicinal of will loss. extinction loss biodiversity the to language critical coupling that more strong suggests even of and languages finding knowledge threatened Our indigenous with languages are. in most knowledge uniqueness threatened, unique high linguis- linguistically not report with are that associated knowledge biocul- one species medicinal unique to high plant 12,495 tically known all most a only of Whereas unique—i.e., have 75% language. linguistically over are that that services regions plant show go we three plants diversity, and on tural much languages how is Focusing as quantify plants vanish extinct. and may medicinal knowledge languages of indigenous individual knowledge with indigenous ask associated loss degree we under- the Here, what our limited. in however, remains to result knowledge far, may unique So extinction be linguistically century. of longer language the no whether will of of world 2021) end standing the 24, the February in review languages by for (received 7,400 spoken 2021 the 9, of May approved 30% and Over AZ, Tempe, University, State Arizona Turner, L. B. by Edited a C Rodrigo unique knowledge of medicinal loss the triggers extinction Language PNAS eateto vltoayBooyadEvrnetlSuis nvriyo uih H85 uih Switzerland Zurich, CH-8057 Zurich, of University Studies, Environmental and Biology Evolutionary of Department deaotpat n hi evcsicuigknowledge services—including their knowl- and sophisticated plants a about accumulated edge have people ndigenous u eut hwta nieoskoldeaotmedicinal about knowledge indigenous that show results Our about knowledge indigenous of structure the Unraveling 01Vl 1 o 4e2103683118 24 No. 118 Vol. 2021 amara-Leret ´ | csse services ecosystem a,1 n od Bascompte Jordi and | iclua diversity biocultural iu insipida Ficus a + fuiu nweg soitdwt hetndlnugsin languages threatened percentage with the Guinea. associated language-threat New underestimate knowledge our likely unique that (15) of suggests Glottolog generation from single data a not con- language do in in mostly who decline in skills dramatic students uses a fluent the Such few languages. in students indigenous species a speak plant to by nonnative known in replaced centrated Guinea uses are of languages New plant 91% Crucially, indigenous medicinal Papua (14). to languages varied in compared indigenous students, in the study fluent 6,190 are recent of parents, 58% their A only assess. that showed to languages Guinea’s New difficult of status unique is island- true an all the of survey, By absence of linguistic the efforts. wide in 31% However, documentation Guinea. for New future in account knowledge and, for languages languages) area threatened threatened priority contrast, most key to and (i.e., a linked America hotspot thus, is that knowledge North knowledge indigenous highlights in an medicinal result are knowledge This Americas unique the respectively. all Amazonia, of northwest 100% and that languages in question found the mostly threatened. raises is are knowledge knowledge America, unique unique North of find- whether pattern in of Our strong 50% respectively. a Guinea, and of differ- ing New 18%, using and 56%, language Amazonia, is northwest each species), as (i.e., species), plant turnover same ent species the to of (i.e., fraction turnover knowledge opposed The cultural different by 2). having explained (Fig. is languages respectively that knowledge language, unique one of only by cited doi:10.1073/pnas.2103683118/-/DCSupplemental at online information supporting contains article This 1 ulse ue8 2021. 8, June Published BY-NC-ND) (CC 4.0 NoDerivatives License distributed under is article access open This Submission.y Direct PNAS a is article This interest.y research; competing performed no paper.y declare the R.C.-L. authors wrote The J.B. research; and designed R.C.-L. and J.B. data; analyzed and R.C.-L. R.C.-L. contributions: Author owo orsodnemyb drse.Eal [email protected] Email: addressed. be may correspondence whom To tn o naeigadcnevn auescnrbtosto Rosetta contributions nature’s knowledge—a people. conserving medicinal and there- unraveling of for is reservoir stone language unique indigenous a Each fore plants. language— threatened than single languages with threatened a with knowledge associated by strongly medicinal more known ecosys- and in most unique—i.e., that such plants linguistically find of of is we example use Here, relevant The services. particularly provide. tem a they is services surrounding medicine their the about and have plants for communities the that importance contain knowledge languages and Indigenous global endangerment development. raise sustainable their to aims about Languages awareness Indigenous of Interna- the Decade as 2022–2032 tional of proclamation Nations United The Significance u nlssidctsta hetndlnugsspot86% support languages threatened that indicates analysis Our https://doi.org/10.1073/pnas.2103683118 . y raieCmosAttribution-NonCommercial- Commons Creative . y https://www.pnas.org/lookup/suppl/ | f5 of 1

SUSTAINABILITY ECOLOGY SCIENCE Fig. 1. Medicinal plant knowledge and its association with indigenous languages. The ﬁgure illustrates a regional pharmacy with remedies (jars with plants) cited by languages (jar labels). In this paper, we assess to what degree the knowledge contained in this pharmacy would be eroded by the extinction of either indigenous languages or plants.

Once we have quantiﬁed the overall amount of unique knowl- three regions (Fig. 3; SI Appendix, Table S1). This indicates that edge, we next proceed by mapping how it is distributed across the when planning for medicinal knowledge conservation, the entire linguistic phylogeny. This will serve to identify whether unique linguist spectrum—rather than a few “hot” branches—needs to knowledge is uniformly distributed across all linguistic groups be considered. or whether a few linguistic groups deserve more protection than So far, we have focused on how unique knowledge is dis- others. First, we built language phylogenies for all of the indige- tributed along the cultural dimension. Let us turn now to nous languages in our sample. Next, we calculated the degree examine the other component of the indigenous knowledge of phylogenetic clustering of unique knowledge using Pagel’s network (9), namely, the plants. To understand the degree of lambda (λ) (16); values of λ close to one indicate strong phy- threat faced by medicinal plants, we queried the International logenetic clustering, whereas values close to zero indicate data Union for Conservation of Nature (IUCN) Red List of Threat- without phylogenetic dependence. We did not ﬁnd clustering of ened Species (17). We found conservation assessments for 22%, unique knowledge along the language phylogenies in any of the 31%, and 32% of the medicinal species recorded in North

Fig. 2. Most medicinal knowledge is unique to a single language. Histograms depict the number of indigenous languages that cite a medicinal service. (A) North America. (B) Northwest Amazonia. (C) New Guinea. Red bars show medicinal plant services only known to one language. Dots within the maps indicate the distribution of languages.

2 of 5 | PNAS Camara-Leret´ and Bascompte https://doi.org/10.1073/pnas.2103683118 Language extinction triggers the loss of unique medicinal knowledge Downloaded by guest on September 29, 2021 Downloaded by guest on September 29, 2021 7, styloptera; Tetrapterys agaeetnto rgestels fuiu eiia knowledge medicinal unique of loss the triggers extinction Language C domain. public 1, region. per services medicinal unique more (n Amazonia northwest i.4. Fig. as change may patterns Meth- observed the and whether northwest (Materials ascertain America, respectively To ods). North Guinea, in New and threatened Amazonia, as and classiﬁed 1%, 4%, were assessments, respectively. IUCN 4% with Guinea, ﬂora 2, medicinal New total Cherokee; the and Of 1, tips Amazonia, trees: phylogeny at linguistic northwest names see the Language America, codes, Molima. in Glottolog 9, and numbers and names Dani; corresponding Valley language their Grand of Lower list by 8, the indicated Biak; For 7, codes. are Cubeo; Glottolog languages 6, following Barasana–Eduria; abbreviated These 5, are region. Ticuna; 4, per Navajo; services 3, Huron–Wyandot; medicinal unique of number (n Amazonia 3. Fig. ABC mr-ee n Bascompte and amara-Leret ´

1245 wa

me sh1 coma1 a1260 e shos1248 j ka w

253 t e

utes123 u w

c t

a1283 a no 45 245 h

127

253 iow1266

u 1248 rt2954 1

1 a yuki1 k

h s 2

itiuino nqekoldears agae.Tesrpeetlnug hlgne fNrhAeia(n America North of phylogenies language represent Trees languages. across knowledge unique of Distribution 8 itiuino nqekoldears eiia oa.Tesrpeetmdcnlpatpyoeiso ot mrc (n America North of phylogenies plant medicinal represent Trees ﬂoras. medicinal across knowledge unique of Distribution opi omah 8 6

sout2961 4 lui m pim toho12 243 124 h an 304 io oc 12 1 5 w d o a1 h124347 14 9 r 271 2 4 ca 124 la 4 6 sh ka ko 5 ta su d as t i1257124840 1 1 a r a 28 yana1 82 k 2 1 3 1248 o 4 se ;adNwGie ( Guinea New and (B); languages) 37 = 1 7 2 mi cr 25 6 hava u 5 o 39 ec1 k h1280 w 8 mari14u o1261 54 1 q t2 kas 244 s 966 coc ea rt2 p o rt2967 awn na n 12 a tc1249 no nezp123854 rik 12 2 uma 62 so t1237 is124 36 ut29 ob 249 c12 85 te1 u ku maka1318 nu yoku1 diti1235 256nort29 heil1246 nort2951 69 69 nise1244 kwak12 hais1244 ;adNwGie (n Guinea New and (B); species) 645 = west2632 quil1240 east1472 quin1251 1 ncru fagifer Inocarpus cher1273 cowl1242 2 uppe14 wyan1247 39 lush1252 arap1274 3 sout gros124 twa 2965 247 clal n12 chey1 g b 1241 47 544 itx ell1 stra ast2 1243 124 hal1244 e h 2 243 sh k pow 129 h 1 u 12 e a th s1 45 wa1245 o 2 mal 2 9 a o m 48 251 24 4 l ka 124 5 1 1 2 e n m t1 0 6 u 1 3 123 uns a 3 e 12 ka 24 m m n n ast2 3 k u nank12412 8 6 li mi ti d or 0 1308 a o 8 5 5 42 zuni t29435 c 2 3 ent wo 126 2 4 1 36 n o 52 ks123 u 47 k1 2 1245 212 c 2 lg s 2 si l mont1 a e ee1270 1 a 7 6 41 m cr 49 ot kla p 12 alab1237 koas1 car ip meno1 c1276 o m1 ch hic1270 h r c c we 1 3 2 me 2

4 5

n1245 49 4 2

st 8, ;

1 u tli sc123

2 a k pp

v o

a tolo125 yu e1

n 3

8 12 4 37

lglai indica Flagellaria 3 2, tulipifera; Liriodendron 7 248 9 1

t uro an y

9 n 0lnugs (C languages) 80 =

c a cu 1 2 4 7 pce)(C species) 477 = 1

co n 9, and ;

f a12

4 yano1 p bo u 2 i n1248 r a1263

262 asha12

urr12

4 c c

l lutain rmwwpatlutain.r eogt the to belong www.plantillustrations.org from illustrations All fruticosa. Cordyline and

abi1241

4 u1

wa 2

ti 2

c or 48 3 u1 1240

3, borbonia; Persea ca 24

r 5 de

.Ilsrtosrpeetidgnu ruswoelnugshv h highest the have languages whose groups indigenous represent Illustrations ). i1279 ni 1241

uru

127

4 a

.Ilsrtosadtercrepnignmesso h ln pce with species plant the show numbers corresponding their and Illustrations ). o 1

2 5

etAaoi,adNwGie,rsetvl.Acrigto According respectively. Guinea, New north- and and America, 25%, North Amazonia, in 57%, recorded west species for medicinal assessments the (Materials of contains (18) 49% which study conserva- Methods), machine-learning obtained a and also from we predictions assessed, tion formally are species more 6

t ani

huam1247 1257

ach

2 4

8 desa1247

1 2 5 7 cu ba be12 6 ra1 42 380 5

macu

inga 1260 41

t tu e

n ca12

a car

san

24 gu

a 52

c 1272

4, glabra; Pinus m an1269

oc ship

2 a1259 si

kuli1255

8 sh

9 on1

1254 seco a

r124 247

2 5

41 al S2. Table Appendix, SI

ka m

waim12 9 ahgl paniculata Tachigali ul m e

si kili1267

12 ke

ang14 wa na12 moli1

suau1242

40 124 mot wa n

m ir o

nd bi 1 a

51 m se

0 or eng1267 24

u1246 r

3 ak1

4 2 12

m 3

ru west2594 west2594 2

67 8

u 38

2 1 8

n 124 26

8 k 44

a1 8 5 w

262 e kuan1248 o

n 1

i126

415

https://doi.org/10.1073/pnas.2103683118 a

n 1

a g k g123

ra a

n1 r12

p 4

2 3 ina1 9

25 kais12337

2 ko

m s a

o uab 12

i 49

i12 1 5

3 2 e

5 i

;northwest (A); languages) 119 = 38 p

k 124 e

ih te12

a 2

a12 5 4

e ng

n a1

ucl 25

1 2

59 h uli

5 12 4

a 4

man1265

b e n a 1 2

pare1271 4

hata1243 fore1270

m a n

i 1

2 3

5 gimi1243

e y

5, ; a 12 3

6 no

rt2920

r a

s 12 ku 49 ma si 1280

na127

6, albivenis; Fittonia m a im

,7 pce)(A); species) 2,475 =

u b

a o

2 1

w 23

m 4

8 f

2 bu u

1 m 262

b y

u e

u r

1 u da k lp 2 1242

47 a 13 y 12

m mo1255 3

f 0 o 8 a cas 8 p

1 a n12

2 moro1289 iw

7 38

2 1 u

2 1

ng tifa1245 237

y o 3 k

e t

k ba

PNAS a12

i e

1 sa12

n bi l

2 e

r mian1

3 n z ucl16

3 1

u 9

5 a12 256 137

cl14 ma124 kal 45 no

5 7 2 2

a139

54 b 7 2 5

o ia

4 1

1254 maib1239 238 7

wari1264

bl12 43

f5 of 3

hamt12

4 7

SUSTAINABILITY ECOLOGY SCIENCE that study, the probability of a medicinal species belonging to from Richard E. Schultes’s book on the medicinal plants of northwest Ama- a threatened category ranged from 0.0002 to 0.8341 in North zonia, which integrates nearly half a century of his field research (12); and America (mean ± SD, 0.156 ± 0.158), 0.149 to 0.822 in north- 3) New Guinea: from an ethnobotanical review of 488 references and 854 west Amazonia (mean 0.483 ± 0.119), and 0.063 to 0.679 in New herbarium specimens (13). Guinea (mean 0.357 ± 0.141), respectively. In summary, both We classified uses from the three data sources into medicinal subcategories following the classification in the Economic Botany Data Collection the IUCN conservation assessments and machine-learning pre- Standard (31), with modifications explained by Camara-Leret´ et al. (32). dictions suggest that most medicinal plant species in our sample Medicinal subcategories included Blood and cardio-vascular system; Cultural are not threatened. Finally, we found that about 1% of all unique diseases and disorders; Dental health; Digestive system; Endocrine system; knowledge in each region was associated with both threatened General ailments with unspecified symptoms; Infections and infestations; languages and threatened plants (SI Appendix, Table S3). How- Metabolic system and nutrition; Muscular-skeletal system; Nervous system ever, there is considerable uncertainty about the potential loss of and mental health; Poisoning; Pregnancy, birth and puerperium; Reproduc- unique knowledge from the extinction of plants because 64% and tive system and reproductive health; Respiratory system; Sensory system; 69% of the unique knowledge in North America and northwest Skin and subcutaneous tissue; Urinary system; Veterinary; Not specified; Amazonia that is associated with threatened languages belongs and Other medicinal uses. We defined “unique knowledge” as a medicinal service cited exclusively by one indigenous language. to plants that lack plant-conservation assessments. IUCN con- The amount of unique knowledge may change as more indigenous servation assessments are urgently needed for these plant groups are studied and as more in-depth studies are made on indige- species. nous groups already reported in the literature. We hypothesize that it will To assess whether unique knowledge is strongly clustered increase, for three reasons. First, research in South America indicates that biologically, we built phylogenies of the medicinal floras of the amount of unique knowledge is positively correlated with the total each region and calculated Pagel’s lambda (Fig. 4). We only number of medicinal uses registered in a community (33). This is confirmed found significant clustering of unique knowledge in North Amer- in New Guinea by a study that showed that linguistic uniqueness occurs ica, although values were low (SI Appendix, Table S1). This even among the best-studied indigenous groups who occupy montane habi- relatively weak phylogenetic signal across the three regions tats with a similar flora (34). Therefore, any undersampling in our study regions—especially in New Guinea—would, in fact, underestimate unique suggests that when planning for biocultural conservation, the knowledge because generalist knowledge tends to be the first that is docu- entire medicinal flora—rather than a few clades—must be mented in the field. Second, our classification of medicinal plant services considered. is conservative because it omits “plant parts” (e.g., bark, leaf, fruit, and Here, we have shown that in North America, northwest Ama- seed). Because different plant parts may be used for different purposes zonia, and New Guinea, indigenous knowledge of medicinal [different plant parts may have different chemical compounds (35)], our plant services exhibits a low redundancy across languages that classification underestimates the detection of medicinal knowledge that is typical of systems with high information content (19, 20). is restricted to one language. Third, our classification of medicinal plant This low redundancy in medicinal knowledge among languages services is based on 20 medicinal subcategories that are broad in scope. does not support the notion of high cross-cultural consensus— For example, “Infections and infestations” encompasses reports related to, e.g., malaria, leishmaniasis, and measles. This means that if one indigenous i.e., that cultures resemble each other in their knowledge—but language cites plant A as a remedy for malaria and another indigenous lan- instead highlights the unique biocultural heritage each culture guage cites plant A for measles, both reports would be classified as “Plant holds. The invention and diversification of languages involve A: Infections and Infestations” and considered shared, rather than unique, two opposing forces. On the one hand, sharing facilitates the knowledge. exchange of information and the spread of valuable ideas that may enhance the fitness within populations. On the other hand, Language Phylogenies and Threat. Medicinal services in the literature were the diversification of languages is the result of innovations, and associated with 119 indigenous languages in North America, 37 lan- eventually linguistic barriers may limit information spread. In guages in northwest Amazonia, and 80 languages in New Guinea. For areas of high linguistic or biological diversity and/or geographic each region, we built language trees through phylogenetic inference barriers, the balance between sharing and innovating may tip using machine-learning techniques on the word lists of the Automated toward the latter. This may result in the amplification of dif- Similarity Judgment Program (ASJP) database (36) and used the Glot- ferences among cultures, as we have shown here for the case of tolog classification as a constraint tree (30). The ASJP list consists of the medicinal knowledge. 40 most stable items, as determined by Holman et al. (37), from the Only about 6% of higher plants have been screened for biolog- 100-item list of Swadesh (38). To assess the degree of threat faced by languages in our sample, we queried the Glottolog (15), which derives ical activity (21). Therefore, assessing to what degree linguisti- an Agglomerated Endangerment Status (AES) from the databases of The cally unique medicinal services are truly effective in the Western Catalog of Endangered Languages, United Nations Educational, Scientific, sense is beyond the scope of this paper. In many instances, these and Cultural Organization Atlas of the World’s Languages in Danger, plants have been proven medicinally effective (12, 22–27), albeit and Ethnologue. For a list of the languages analyzed, see SI Appendix, there are also exceptions (28, 29). Regardless of that, here, we Table S2. treat this knowledge as what it is: part of the cultural heritage of indigenous people. Vascular Plant Phylogenies and Threat. We verified plant-species taxonomy The United Nations declared 2022–2032 as the International using recently published checklists to the vascular plants of the Americas Decade of Indigenous Languages to raise awareness about their (39) and New Guinea (40). Using the list of medicinal plant species in each importance for sustainable development and their endanger- region, we queried the mega-tree GBOTB.-extended of Smith and Brown ment across the world. Our study suggests that each indigenous (41) with the phylo.maker function of the R package V.PhyloMaker (42). language brings unique insights that may be complementary to The phylogenies used in all subsequent analyses comprised 2,475 species other societies that seek potentially useful medicinal remedies. in North America, 645 species in northwest Amazonia, and 477 species in Therefore, the predicted extinction of up to 30% of indigenous New Guinea. To assess the threat faced by medicinal plant species, we queried the conservation assessments published by the IUCN Red List of languages by the end of the 21st century (3) would substantially Threatened Species (17), which classifies species as Data Deficient, Least compromise humanity’s capacity for medicinal discovery. Concern, Near Threatened, Vulnerable, Endangered, Critically Endangered, Extinct in the Wild, and Extinct. Following IUCN, species assessed to be Materials and Methods Near Threatened, Vulnerable, and Endangered were considered threat- Plant Services. We obtained a list of medicinal plant species and services ened. Because most plant species lack IUCN conservation assessments, associated with individual indigenous groups from three regions: 1) North we also obtained endangerment probabilities from a recent study that America: from the Native American Ethnobotany database (11)—the largest used machine-learning to predict the conservation status of 30,497 plant repository of indigenous knowledge for the region; 2) northwest Amazonia: species (18).

4 of 5 | PNAS Camara-Leret´ and Bascompte https://doi.org/10.1073/pnas.2103683118 Language extinction triggers the loss of unique medicinal knowledge Downloaded by guest on September 29, 2021 Downloaded by guest on September 29, 2021 1 .Vrore hraons ntenwmlenu:Lafidn n biotechnol- and Leadfinding millennium: new the in Pharmacognosy Verpoorte, biology. R. in theory 21. Information Margalef, R. communications. 20. of theory mathematical A Shannon, E. C. 19. 8 .A elte,B .Crtn,D .Tn,J ulvn .Esp A. Sullivan, J. Tank, C. D. Carstens, C. threatened B. of Pelletier, list A. T. red 18. IUCN The Nature. of Conservation for Union International 17. evolution. biological of patterns historical the Inferring Pagel, Mark. 16. 5 .Hammarstr H. 15. Kik A. 14. C R. 13. dyes Raffauf, F. drugs, R. foods, Schultes, E. of R. database 12. A ethnobotany: American Native Moerman, D. 11. Nettle D. 10. agaeetnto rgestels fuiu eiia knowledge medicinal unique of loss the triggers extinction Language C ltoo 1) osrainassmnsfo UN(7,adconservation (18). and al. et (17), Pelletier IUCN from from predictions assessments conservation (15), J Glottolog from available are trees Language 13). Availability. Data .R C R. 9. Pironon S. 8. C R. 7. knowledge ecological local Saynes-V of A. trends Global 6. Sauer, and H. languages H. W. of Lemahieu, distribution A. Aswani, global S. and 5. risk extinction in Parallel Sutherland, update” J. 20-year W. A 4. crisis: in languages world’s “The Lewis, P. M. Simons, F. G. 3. Berlin, B. 2. McDade W. T. 1. mr-ee n Bascompte and amara-Leret ´ ogy. (1948). 423 (2018). scale. global a on priorities conservation pce Vrin22-,ItrainlUinfrCnevto fNtr,Cambridge, Nature, of Conservation 2020). for UK, Union International 2020-2, (Version species (1999). 884 2020. February 1 Accessed https://glottolog.org/. review. A change. global of face change. climate Adv. Sci. Mexico. of Zapotecs Isthmus the Ethnomed. among knowledge ethnoecological implications. future and Amsterdam, Benjamins, John species. Series, Companion 3–19. pp. Language 142, vol. in 2013), Netherlands, (Studies Eds. ley, Endangerment Language to Responses Societies Traditional 2014). in Animals and Amazon. Bolivian the in health 6Arl2021). April (Accessed 26 https://doi.org/10.1101/2021.04.12.439439 (2021). [Preprint] bioRxiv threat. Amazonia naeb.brit.org/. Northwest (2020). plants from derived 2020. peoples, May American 24 Accessed Native of fibers and 2000). UK, Oxford, Demand, on Press University mr-ee,M .Frua .Bsope nieoskoldentok nthe in networks knowledge Indigenous Bascompte, J. Fortuna, A. M. amara-Leret, amara-Leret mr-ee,Z enh,Idgnu nweg fNwGie’ sflplants: useful Guinea’s New of knowledge Indigenous Dennehy, Z. amara-Leret, ´ ´ ´ .Pam Pharmacol. Pharm. J. h ol’ oso flnusi n iclua iest under diversity biocultural and linguistic of hotspot world’s The al., et Nature az45(2019). eaaz1455 5, tal., et cn Bot. Econ. oeta dpiesrtge o 9sbShrncosudrfuture under crops sub-Saharan 29 for strategies adaptive Potential al., et tnbooia lsicto:Picpe fCtgrzto fPlants of Categorization of Principles Classification: Ethnobiological 0(2013). 40 9, suz .Cbleo .A ev,F hag utrlcag n osof loss and change Cultural Chiang, F. Meave, A. J. Caballero, J. asquez, ´ tnbtnclkoldei soitdwt nie fchild of indices with associated is knowledge Ethnobotanical al., et m .Fre,M aplah .Bn,Gotlg43(2021). 4.3 Glottolog Bank, S. Haspelmath, M. Forkel, R. om, ¨ 7–7 (2003). 276–279 423, a.Ci.Change Clim. Nat. aihn ocs h xicino h ol’ Languages World’s the of Extinction The Voices: Vanishing lmt hnetraesNwGie’ iclua heritage. biocultural Guinea’s New threatens change Climate al., et aao eiia ln evcsaepbil vial (11– available publicly are services plant medicinal on Data 0–1 (2019). 405–415 75, DocrdsPes otad R 1990). OR, Portland, Press, (Dioscorides rc al cd c.U.S.A. Sci. Acad. Natl. Proc. lSOne PloS 5–6 (2000). 253–262 52, h eln oet eiia n oi lnso the of Plants Toxic and Medicinal Forest: Healing The rc al cd c.U.S.A. Sci. Acad. Natl. Proc. 0940(2018). e0195440 13, 5–6 (2019). 758–763 9, .Mhs .Pre,G e-oa,K Wheat- K. Rei-Doval, G. Perley, B. Mihas, E. . rc al cd c.U.S.A. Sci. Acad. Natl. Proc. PictnUiest rs,Pictn NJ, Princeton, Press, University (Princeton gr(0,lnug hetdt from data threat language (30), ager ¨ e.Syst. Gen. 9391 (2019). 9913–9918 116, 67 (1958). 36–71 3, 1463 (2007). 6134–6139 104, elSs.Tc.J Tech. Syst. Bell noa rdcigplant Predicting ´ ındola, 13027–13032 115, Nature .Ethnobiol. J. 877– 401, 379– 27, (Oxford 2 .Jn .Qa,VPyoae:A akg htcngnrt eylrephylogenies large very generate can that package R An V.Phylomaker: Qian, H. Jin, Y. 42. 1 .A mt,J .Bon osrcigabodyicuiese ln phylogeny. plant seed inclusive broadly a Constructing Brown, W. J. Smith, A. S. 41. 0 .C R. 40. Ulloa U. C. dating. lexicostatistic 39. in accuracy greater Towards Swadesh, M. 38. 2016) 17, (Version database Holman W. ASJP E. The Eds., 37. Brown, H. C. Holman, W. E. Wichmann, S. 36. J G. 30. 5 .F afu,Sm oe ntedsrbto faklisi h ln kingdom. plant the in alkaloids of distribution the on notes Some Raffauf, F. R. 35. C R. 34. C R. 33. C R. 32. Cook M. E. F. 31. Guinea New A ethnomedicine: in ethnobotany Applequist medical L. W. of 29. role The Telban, B. 28. Prescott K. A. T. Papua of 27. plants Medicinal Simmonds, J. S. M. Kiapranis, R. Briggs, M. Prescott, K. A. T. 26. therapeutics. new Svetaz, of L. sources as 25. plants Medicinal Elvin-Lewis, P. M. Lewis, H. W. the 24. from agents therapeutic of identification the and “Ethnobotany Balick, Samoan J. the M. of activity Pharmacological 23. Bohlin, L. Tuominen, M. Sperry, R. L. Cox, A. P. 22. ee o h lutain nFg.1ad3 hswr a enspotdby supported been has C work I. This 310030 and 3. Grant and discussions, Foundation 1 Science helpful Figs. National for in Swiss illustrations laboratory the J.B. for the Leret of members the ACKNOWLEDGMENTS. o aclrplants. vascular for .Bot. J. (2020). 579–583 Americas. (1955). 121–137 (2008). 331–354 2020. April 14 Accessed https://zenodo.org/record/3835942. Bot. assessments. science-policy for prop- intellectual assessing America. South (2014). in northwest 58–65 in serve benefit-sharing knowledge and rights traditional erty in patterns Geospatial America. South (2014). northwestern e85794 in under-documented vastly is knowledge 1995). UK, Kew, Gardens, Data bioassays. antimalarial in performance example. nomads. Apsokok lime plant–slaked of use their on focus a mixtures. and countries. population speaking American Miu Guinea’s Latin New seven among survey A Ethnopharmacol. properties. antifungal with 154, vol. 1990), Gard., Bot. UK, Mo. Chichester, Sons, and Wiley 22–39. pp. Symposium, Foundation (Ciba in rainforest” ethnopharmacopoeia. gr lblsaepyoeei igitcifrnefo eia resources. lexical from inference linguistic phylogenetic Global-scale ager, amara-Leret mr-ee,Z enh,Ifraingp nidgnu n oa knowledge local and indigenous in gaps Information Dennehy, Z. amara-Leret, ¨ ´ mr-ee,N aigaZmrn,H ase,M .Mac J. M. Balslev, H. Paniagua-Zambrana, N. amara-Leret, ´ mr-ee,N aigaZmrn,J-.Senn,H ase,M .Mac J. M. Balslev, H. Svenning, J.-C. Paniagua-Zambrana, N. amara-Leret, ´ 43 (1970). 34–38 24, ´ –6(2018). 1–16 5, 0–1 (2018). 302–314 105, .Ethnobiol. J. .Ethnopharmacol. J. Science au fteehoeia nomto o h icvr fplants of discovery the for information ethnomedical the of Value al., et nitgae seseto h aclrpatseiso the of species plant vascular the of assessment integrated An al., et xlrtosi uoae agaeclassification. language automated in Explorations al., et tal., et .J hdik .Mrh Eds. Marsh, J. Chadwick, J. D. Plants, from Compounds Bioactive 62 (1995). 16–24 82, rpclucrpattetet sdb au e Guinea’s New Papua by used treatments plant ulcer Tropical al., et e unahstewrdsrcetiln flora. island richest world’s the has Guinea New al., et niaailueo aaaypat spol orltdwith correlated poorly is plants Malagasy of use Antimalarial al., et 3–5 (2010). 137–158 127, 6411 (2017). 1614–1617 358, .Ethnopharmacol. J. Ecography cnmcBtn aaCleto Standard Collection Data Botany Economic 4–6 (1988). 149–169 8, cn Bot. Econ. etakG J G. thank We 1–2 (2015). 217–223 174, 3315 (2019). 1353–1359 42, a.Sustain. Nat. 8–9 (1989). 487–497 43, 4–4 (2017). 240–245 205, cn Bot. Econ. grfrpoiigtelnug trees, language the providing for ager ¨ https://doi.org/10.1073/pnas.2103683118 3–4 (2019). 736–741 2, 58 (2017). 75–82 71, 921(oJ.B.). (to 197201 .Ethnopharmacol. J. n.J m Ling. Am. J. Int. a Ethnobotanical ´ ıa, PNAS RylBotanic (Royal oi Ling. Folia Nature lSOne PloS | amara ´ f5 of 5 Econ. Ann. 584, 158, Am. Sci. 42, 21, ´ ıa, 9, J.

SUSTAINABILITY ECOLOGY SCIENCE