Simplex Generic Toponyms in Four English-speaking Jurisdictions

Toponymic literature generally recognizes three main structures in English: those with a specific + generic element (e.g. Tweed River, Mount White); those that consist of a specific element alone (e.g. Perth, Washington); and less commonly, those with a generic element preceded by the definite article (e.g. The Bend, The Bluff). There are of course combinations and variations on these three (e.g. The Maiden Mountain, Valley of the Giants). Very rarely has the existence of toponyms comprising solely a generic term (e.g. Pinnacle, Sugarloaf) been mentioned or discussed. This survey investigates the occurrence and use of such toponyms in four English-speaking regions.

etc.) to the class of proper names" (ICOS n.d.,4). Using a similar term, but specifically referring to the process in toponyms, UNGEGN (Kadmon 2007, 4) refers to it as "toponymization", i.e. "The act of producing a toponym from a common noun or other part of speech." Room (1996, 10) refers to the process as the "appellative stage", that is, "the creation of a name from an appellative [which] usually applies when the form or meaning of the name is already present in the appellative, and when the derivation of the name is onymic." By "appellative," Room and Harval ık mean "a common noun, or phrase denoting a number of identical objects, as distinct from a name, that denotes only one." (Room 1996, 10). In other words, a proper name uniquely and specifically identifies an individual entity, whilst a common noun designates any one of a particular entity or class. There is therefore, a binary opposition in the onymic and appellative aspects of language.
Proprialisation, in many languages, is not manifested by changes in the structure of the original appellative. In English, for example, the newly formed proper name and the appellative are formally identical, with the exception of the first letter being uppercase for proper names. In Slavic languages though, the difference between homonymous appellatives and proper names is indicated by a change in gender and the declension paradigm (Harval ık 2012, 11-12). In any case, it is usually evident from the context of its use whether the term being employed is an appellative or a proper name. Moreover, as soon as an appellative becomes a proper name, it behaves accordingly, observing the rules of the onymic sphere of language.

Methodology
The ensuing survey examines and analyzes SGTs in four distinct English-speaking territories: Australia (AU) Canada (CA), New Zealand (NZ), and the United States (US). The reason for choosing these four regions is that they have easily accessible national gazetteers, and that they are all English-speaking countries. Comparisons made between them will increase the likelihood sensible and valid parallels to be drawn, should such exist.
The online gazetteers of Australia (Geoscience Australia 2016a), Canada (Natural Resources Canada), New Zealand (Land Information New Zealand), and the United States (United States Geological Survey) were used to gather the data for this study. In order to conduct a valid comparison between the jurisdictions, only English SGTs were recorded because the indigenous languages of these regions are too numerous and diverse to draw any meaningful parallels in this instance. The sources used to compile an index of search terms for the gazetteers are: AU: Glossary of Generic Terms (Geoscience Australia 2016b); Blair (2014); Blair and Tent (2015) CA: Geographical Feature Type (Natural Resources Canada) NZ: Generic Geographic Features Listing (Land Information New Zealand) US: Feature Class Definitions (U.S. Geological Survey) Because SGTs referring to non-natural geographic features were found to designate just civic and constructed features, only terms for natural geographic features were included in the index. 4 The focus of the current survey was to determine how many, if any, SGTs defined natural features. I was particularly interested those SGTs that had retained the original or literal designations for those natural features. This would aid in an understanding of how GSTs come about, and what kinds of natural features were prone to obtaining such a toponymic form. SGTs referring to non-natural features were considered "false generics." It must be noted that the geomorphology of each jurisdiction has a bearing on the type of GFTs employed. For instance, NZ and the US have numerous fumaroles and geysers (vents in active volcanic areas from which steam, super-heated water, gases and various acids are ejected) (e.g. Lady Knox Geyser and Old Faithful Geyser respectively), and glaciers (e.g. Franz Josef Glacier and Andrews Glacier respectively). In Australia, none of these are found.
Finally, jurisdictions may classify GFTs under dissimilar feature classes or define them differently. 5 For instance, Geoscience Australia's feature classes are more particularized than those of the USGS-the AU classification distinguishing between HILLs, MOUNTAINs (MTs), and PEAKs, whereas the US schema classifies them all under SUMMITs (SUMs). Table 1 itemizes some examples where the AU and the US feature classes are defined and classified differently. Table 1 reveals substantial differences not only between the USGS and Geoscience Australia's feature classes but also their GFTs. The Australian system has more feature classes, whilst in some instances the USGS has more GFTs. The latter being most likely due to the different geomorphologies of the two regions. The USGS GFTs also contain terms from Spanish and Native American languages. 6 Where a discrepancy exists between either the meanings of a generic feature class or in the GFTs employed between jurisdictions, a synonymous or suitable alternative term was found when classifying the toponyms in the dataset. Unless, a feature class was unique for a jurisdiction and essential for the sake of clarity, the terms in the Feature Class Definitions for the USA have been employed in this survey.

Results
The total number of distinct GFTs that have been proprialized and used as SGTs in the four regions was a mere 144. As Table 4 shows, a number of these are duplicated as toponyms (SGTs) for various places, but overall, the total number in the four regions surveyed is still very small. Table 2 shows that such toponyms comprise only a tiny proportion (between 0.03% and 0.13%-with a mean of 0.08%) of the total number of gazetted toponyms in each jurisdiction, with AU and NZ having the highest proportion of them. Table 3 summarizes in more detail the results obtained from the data collected. It shows the total number of SGTs in each of the four regions, and the type of geographic feature named (i.e. natural vs. non-natural). Natural features are further divided into those that reflect or express the literal meaning of the proprialised GFT (e.g. AU Reef > REEF; US 112x Sugarloaf $ Sugar Loaf > SUM), and those that express a non-literal meaning of the proprialised GFT (e.g. CA Glacier > WRFL 'waterfall'; NZ Washpool > HILL). As can be seen, the results do not reveal any overt patterns.
The vast majority of SGTs are found in the US (1069 À 61.6% of the 1736 total in the four regions), just over twice as many as Australia's 508 (29.3%), and more than ten times as many as in Canada (5.6%) and New Zealand (3.5%) respectively. Numerically, the US has the most named non-natural features (842 À 79% of its 1069), however, proportionately Australia has slightly more (408 À 80% of its 508), leaving Canada and New Zealand each with only 10% of such toponyms. Once again, there is no overall discernible pattern to be found in the data as to the referents of the SGTs across the four jurisdictions. Table 3 also shows the US and AU have the most comparable patterns of SGTs. They have the lowest percentage of non-literal meanings, a comparable percentage of literal meanings, and the most designating non-natural features. Given the fact that only SGTs denoting natural geographic features were the main motivating factor for the survey, 7 it is interesting to see that overall nearly 73% of them actually refer to non-natural features, leaving just over 27% that refer to natural ones. Almost all the non-natural features (e.g. POPLs) in the dataset obtained their names from an adjoining or nearby natural feature bearing that name. Escarpment (POPL) near Niagara Falls, and Estuary (POPL) at the port of Tampa (FL), are clear examples of this phenomenon. Others are the result of a ellipsis of the original name. For instance, the Western Australian town of Boulder is an ellipsis of The Great Boulder, the original name for the gold mining lease, the main features of which were large sandstone boulders in which gold veins were found (Casey & Mayman 1964). Similarly, the name of Inlet (POPL) in the state of New York (US) is also derived from a former longer name, Inlet on Fourth Lake (United States Geological Survey). Often names of well-known geographic features are also truncated. This is sometimes exhibited with mountains, e.g. Everest, Kilimanjaro, Matterhorn, Erebus, Kosciuszko, Ruapehu, Tarawera etc. Such constructs are suggestive of hypocoristic names, or the personification of such features. Other prominent or conspicuous features also experience such name pruning, but only when the truncated name cannot be confused with another feature, e.g. Old Faithful (Geyser), (Lake) Eucumbene, Sow and Pigs (Reef), Kakadu (National Park) etc. The ellipsis in these instances results in what will be termed here as a "simplex specific toponym" (SST).
Other truncated names may be derived from GFTs that form the specific element of toponyms inspired through metaphoric analogy because the feature  Table 4 shows that two thirds (66.3%) of SGTs embody GFTs for non-natural geographic features (n. 537: 273). However, what is noteworthy is that 22 (73.3%) of the 30 most common SGTs refer to hypsographic feature terms (i.e. for relief features), 16 of which (72.7%) represent orographic feature terms (i.e. for elevated features). It seems therefore that orographic features tend to produce more SGTs than most other GFTs. This seems reasonable because it is precisely these elevated features that stand out in the surrounding landscape from other feature types. The remaining eight SGTs exemplify vegetation feature types (n. 4 or 13.3%) and water feature types (n. 4 or 13.3%). Of the orographic features, Sugarloaf$Sugar Loaf produces the most common SGT, with 174 instances across the jurisdictions AU, CA and US. Hillside is the next most popular with 151 instances across AU and US.
Why Sugarloaf is the most common SGT for any natural geographic feature in the dataset (51x SUM in AU; 11x LAKE in CA; and 112x SUM in the US) seems intriguing. Notwithstanding the enigma of the 11 lakes bearing that moniker in CA, the 163 instances in AU and the US designating summits stands to reason. In a forthcoming paper, I attempt to answer why Sugarloaf is such a popular SGT (see in Tent 2020, forthcoming). Table 5 displays the most common feature classes of SGTs in each jurisdiction. The most common ones being for non-natural features (n. 1236 or 76.3%); among them 651 POPLs, 184 LOCs, and in the US, and 226 HMSDs in AU. This is interesting given the original appellatives upon which these SGTs are based all denote natural geographic features. The link between them is revived if there is some kind of connection between the named object and the appellative whose name forms are homonymous.
Hillside, Sugarloaf e. bimorphemic (open compound-free morpheme & free morpheme), e.g. Jump Up, Sugar Loaf f. a combination of (d)/(e) þ (b), e.g. Tablelands The majority (n. 115 or 80%) of the 144 distinct proprialised appellatives are types (a), (b) or (c), the other 20% (n ¼ 29) are of types (d), (e) and (f). These latter types are analogous in form to toponyms with standard SPECIFIC þ GENERIC structures, either in solid or open compound forms, such as Rutherglen, Bankstown, Forestville, Mossvale, and Lane Cove, Violet Town, Moss Vale, etc. It is perhaps because of this analogous SPECIFIC þ GENERIC structure that such forms are disposed to become proprialised. This, however, still leaves the question as to why types (a), (b) and (c) are more often proprialised. Legend: IS 'island', REGN 'region', STRM 'stream', TRIG 'triangulation station/pillar'.

Discussion & Conclusion
The literature on the nature of proper names is extensive (see van Langendonck 2007 for a comprehensive bibliography). However, very few authors have directly addressed the phenomenon of SGTs or any other kind of simplex generic name (e.g. Dog the name of John Wayne's dog in the 1971 movie "Big Jake"). Kepsu (1997, 118) for instance, lists types of vicinity names, the first of which he labels "topographical terms (location)", and although the examples he provides are Finnish settlement names, they have their origins in GFTs, e.g. Mummi and Malm 'heath', Lahti and Vik 'bay', and Kullo 'hill'. Kadmon (2000, 41) also briefly mentions their existence, although my research shows he has somewhat overestimated their frequency: "In a not inconsiderable number of cases a generic term itself constitutes a toponymy, oftenthough not always-of simplex form." And Anderson (2007, 309-310) comes close to it when he discusses 'generic names' (i.e. names that have pragmatic generic reference) when he states: "Genericness is apparently at odds with the individualization associated with names," which he suggests are based on [common] nouns. However, it is perhaps not surprising that SGTs and simplex generic names (in the sense I use the term in this article), have largely eschewed examination in onomastic literature, perhaps due to their general scarcity. Matthews (2018) does discuss SGTs in some, though not great, detail. In a paper investigating New Zealand's toponyms, using a revised version of UNGEGN's inferred four-tiered model for toponyms (see Kadmon 2002Kadmon , 2007, he argues the UNGEGN model has several short-comings. One is that UNGEGN does not recognize "generic element simplex geonames consist[ing] of one or more morphemes or words that form only a generic element" (Matthews 2018, 101), in other words, SGTs. Matthews schematizes the inferred UNGEGN model as in Figure 1, and shows his revised and expanded model as in Figure 2: The left-hand branch of the "terrestrial geoname [geographical name]" node or member of Figure 2, has one member, viz. "simplex geoname", which in turn has two leaves of the same rank (i.e. peers), viz. "with only a specific element" and "with only a generic element". The latter acknowledges the existence of SGTs as described above.
Matthews also includes in his 'simplex geoname with only a generic element' category items such as The Basin, The Bight, etc. If my contention (and that of Zinkin 1969), as articulated above, concerning such toponyms is accepted, these forms cannot be considered as simplex given they consist of two constituents (a GENERIC and a preposed definite article functioning as a SPECIFIC). In support of this argument, we see that UNGEGN (Kadmon 2002, 18, 19) distinguishes between "composite" and "simplex names" the former consisting of a generic and specific element, or of a specific element consisting of more than one word, and the latter of a specific component only. In addition, Room (1996, 25, 92) differentiates between "composite names" (one-word names comprising at least two morphemes); "compound names" (consisting of at least two components or

SIMPLEX GENERIC TOPONYMS IN FOUR ENGLISH-SPEAKING JURISDICTIONS
two separate hyphenated words); and "simplex names" (consisting of a single component). Lastly, van Langendonck and van de Velde (2016, 34) categorize toponyms into four sets: those with "zero marking" (London, Berlin); those with "suffixing" (Fin-land, German-y); those with "preposed articles" (the Highlands, The Rhine), and those with a "classifier", and possibly an article (the North Sea, the Gobi Desert). It does not appear that toponyms like The Basin and The Bight are classified as "simplex" in any of other toponymic literature. Notwithstanding this minor technical disparity between Matthews' inclusion of GFTs with toponyms with preposed articles, he nevertheless recognizes that SGTs form as a distinct class of toponym, which makes for a useful contribution to toponymic studies.
If SGTs are found in AU, CA, NZ and the US, it is only reasonable to assume that they also occur in other regions of the world and languages. A cursory examination of a number of published articles, gazetteers, and placename dictionaries bears this out, e.g. Gazetteer of British Place Names (Association of British Counties), The Concise Oxford Dictionary of English Place-Names (Ekwall 1947), Dictionary of Southern African Place Names (Raper et al. 2014), Nederlandse Plaatsnamen-Herkomst en Historie ['Dutch Placenames-Origins and History'] (van Berkel & Samplonius 2006), NIS Gazetteer -France (CIA 1964), Places in the World-France (Markowski n.d.), Tongan Place Names (Gifford 1923), New Zealand Gazetteer (Land Information New Zealand), and Matthews (2018). As with many of the English toponyms in AU, CA, NZ and US, many of the SGTs in the UK, South Africa, the Netherlands, France, and Tonga, are LOCs or POPLs, in other words, habitative toponyms. No doubt many of these will have taken their names from neighboring natural geographic features, and in some instances are likely to be the result of an ellipsis of their original name forms. Clearly, a more systematic investigation is warranted to examine SGTs in other countries.
What this small survey has shown is that SGTs constitute a distinct class of toponym not only in English-speaking regions, but also in other regions and languages. They seem to be quite universal, and hence, deserve not only to be recognized as a discrete toponym class, but also warrant further investigation.

Notes
1. I draw a distinction between (a) "generic term", i.e. a common noun designating a type of topographic feature, and (b) "generic element", i.e. that part of a toponym that consists of a generic term. A "generic term" should be viewed in the same light as a common noun, a lexical item that can be defined grammatically or morphologically. By comparison, a "generic element" should be seen as an element of a  Quirk et al. (1985, 265 ff.) distinguish between "specific" and "generic reference" when dealing with article use in English. In the former, a particular specimen, individual or object is being referred to; in the latter, no particular reference to a specific individual or object is signified.  (Kadmon 2002, 7). Feature classes operate at a higher level of abstraction than GFTs themselves, and each is labelled with an alpha code of 2-5 characters. Most can be manifested by more than one GFT, e.g. backwash, brook, burn, creek, river, rivulet, run, etc. come under the feature class STRM. 6. None of which were found to comprise SGTs in the US. 7. Terms for civic features such as: canal, dam, mine, park reserve, tower, municipality, parish, province are excluded.