Understanding Location Descriptions in the LEI System 
David N. Chin Matthew McGranaghan Tung-Tse Chen 
Dept. of Info. & Computer Sciences Dept. of Geography Dept. of Info. & Computer Sciences 
University of Hawaii University of Hawaii University of Hawaii 
2565 The Mall 2424 Maile Way, Porteus Hall 445 2565 The Mall 
Honolulu, HI 96822 Honolulu, HI 96822 Honolulu, HI 96822 
Chin@Hawaii .Edu matt@uhunix, uhcc. hawaii .edu 
Abstract 
Biological specimens have historically been la- 
beled with English descriptions of the location 
of collection. To perform spatial, statistical, 
or historic studies, these descriptions must be 
converted into geodetic coordinates. A study 
of the sublanguage used in the descriptions 
shows much less frequent than typical usage of 
observer-relative relations such as "left of," but 
shows problems with name ambiguity, finding 
the referents of generic terms like "the stream," 
ordinal numbering of river forks and valley 
branches, object-oriented prepositions ("be- 
hind"), fuzzy boundaries (how close is "at," 
how far is still "north of"), etc. The LEI system 
implements a semi-automated understanding 
of such location descriptions. Components of 
LEI include a language analyzer, a geograph- 
ical reasoner, an object-oriented geographic 
knowledge base derived from US Geological 
Survey digital maps with user input, and a 
graphical user interface. LEI parses preposi- 
tional phrases into spatial relations, converts 
these into areas, then computes polygon over- 
lays to find the intersection, and returns the 
minimum bounding rectangle. The user is con- 
suited on unknown words/phrases and ambigu- 
ous descriptions. 
1 Introduction 
Many biological specimens collected in the past 1 are la- 
beled with only an English description of their location 
of collection. 2 To perform any statistical or spatial anal- 
ysis of this historical data, these descriptions must be 
converted into geodetic coordinates (latitude-longitude 
or UTM), a time-consuming process that requires eye- 
straining poring over maps to search for each location. 
1Current collectors can use hand-held satellite-based geo- 
positioning systems to record collection coordinates. 
2There are an estimated several hundred millions of such 
labeled specimens. 
To automate this process requires understanding the nat- 
ural language descriptions, reasoning about the spatial re- 
lations described by the natural language, and mapping 
these into a geographical object base to derive the collec- 
tion coordinates. 
2 Previous Research 
Talmy \[1983\], Herskovits \[1986\], and Andr6 etal. \[1986\] 
among others have documented the many problems in in- 
terpreting and using spatial prepositions. For example, in 
and on have similar but different meanings: "in the car" 
means within the car, while "on the car," means on top of 
the car. However, "on the bus/plane," means within the 
bus or plane. Also, each preposition typically has several 
different meanings or usages. For example, one says "at 
home," but "at the bank," and the meaning of "the plane 
is at Honolulu airport," is within the area of Honolulu air- 
port, but the meaning of "the dog is at the telephone pole" 
is not within the telephone pole, but near it. These context 
dependent usages make interpretation and application of 
spatial prepositions problematic. 
Kuipers \[1985\], Davis \[1986\], Peuquet and Ci-Xiang 
\[1987\], and Frank \[1991\] have investigated qualitative 
and/or quantitative reasoning techniques for dealing with 
spatial relations. Freeman \[1975\] and Mark and Frank 
\[1991\] have identified commonly used spatial relations. 
3 Characteristics of the Domain 
Although general purpose natural language processing 
(NLP) is beyond current state-of-the-art, limited domains 
have frequently been amenable to NLP using specific 
techniques because the domains use a "sublanguage," 
a fairly restricted subset of a general natural language, 
which may have its own syntax and peculiarities. In this 
case, an analysis of one thousand three hundred and forty 
sample location descriptions from the Bishop Museum's 
Herbarium Pacificum collection (accumulated by about 
two hundred different collectors over a period of 160 
years) shows a highly restricted use of language that is 
amenable to understanding using specialized techniques. 
Because these descriptions are meant to be read later 
by a reader who is not at the site, they contain very few 
observer relative descriptions (e.g., behind). Also, there 
138 
are limits to the scale of the descriptions. For example, 
out of a thousand descriptions that were located manually, 
about half of the descriptions were judged to be accurate 
to within 1/3 of a mile and 73% accurate to within 1 mile. 
At the other end, there were no descriptions with accu- 
racy in the meter range and the best descriptions were 
only accurate to within several tens of meters. 
A typical location description is: "Punaluu Valley; 
Castle Trail from Punaluu to Kaluanui Valley + stream, 
on E. side of Northern Koolaus." Associated acces- 
sion information typically includes the date, collector 
name(s), genus, elevation, the museum's collection num- 
ber, and the collector's accession number. This sublan- 
guage is made up mostly of named objects (e.g., "Punaluu 
Valley" and "Castle Trail") and prepositional phrases 
(e.g., from Punaluu to Kaluanui Valley + stream, on E. 
side of Northern Koolaus). The relation of the collection 
location to the unmodified named objects is almost al- 
ways "within," that is, the collection location is within the 
area designated by the named geographic object. The in- 
terpretation of the prepositions is somewhat simpler than 
the general case because the sublanguage deals only with 
a two dimensional cartographic space supplemented by 
elevation markings. This sublanguage is relatively sim- 
ple syntactically, but there are still many problems for 
automatic interpretation of the sublanguage. 
One of the most common problems in interpreting this 
sublanguage is the inconsistent use of names. For exam- 
ple, Waikane-Schofield Trail also appears as Schofield- 
Waikane Trail, Schofield-Waikane Ditch Trail, Schofield 
Trail, W-S Ditch Trail, Schofield Waikane Trail, and 
Waikane Ditch Trail. A mountain like Kaala may be re- 
ferred to as Mt Kaala, Mt. Kaala, Kaala Mrs., Kaala 
Mountain(s), Mount Kaala, Kaala Puu (puu is the Hawai- 
ian word for mountain), Kaala Summit, or Kaala Range. 
Another problem is that names are often not unique. For 
example, Manoa is the name for Manoa Falls, Manoa 
Valley, Manoa Valley Park, Manoa Triangle Park, Manoa 
Stream, Manoa Elementary School, Manoa Japanese 
Language School, Manoa Tunnel, and Manoa Falls. 
When Manoa appears by itself, which Manoa is meant 
is usually clear from the context. In many cases, the 
same name is even used for the same type of object (e.g., 
many cities have Elm Streets and Main Streets). Luckily, 
similar objects with shared names tend to be geograph- 
ically separated, otherwise confusion would result. An- 
other very frequent problem is missing names. Collec- 
tors often use generic terms like stream and gulch to re- 
fer to landmarks that do have names. As before, the con- 
text is usually enough to find the correct object reference 
even without knowing the name. This heavy reliance on 
context for disambiguation is also a frequent problem for 
general purpose NLP systems. 
A difficult problem is the interpretation of ordinal 
numberings, which are used to differentiate forks of 
streams and branches of valleys. For example, "3rd 
branch S of S fork of Kahanaiki Stream," refers to the 
third branch after the main South fork of Kahanaiki 
Stream counting from the head of the stream. Unfortu- 
nately, this description could also refer to the 3rd branch 
following the path of the collector going up the stream. 
Similarly, "Honolulu Valley, 4th small gulch," refers to 
4th small gulch counting from the open head of the valley, 
although this could easily be interpreted as the 4th small 
gulch along some trail that might enter the valley from 
some pass over the mountains at the tail end of the val- 
ley. Another problem is the occasional use of land cover- 
age types such as "middle Metrosideros forest," "wooded 
gulch," and "Fern Forest." Not only do most geograph- 
ical databases lack land coverage information, but such 
information changes frequently over time. Also, descrip- 
tions sometimes refer to rainfall frequency, sun exposure 
or other ephemeral attributes of the area: "in dryish for- 
est," "wet valley," "deep shade in wet gulch," and "shady 
hillside." 
Even after converting the location descriptions into the 
appropriate spatial relations, there are still many prob- 
lems in the correct interpretation of the relations. For 
example, "along a stream" does not mean that the col- 
lection site was in the stream, but within some distance 
of the stream. The problem is what exactly is the value 
of that distance. Even cardinal directions like "north of" 
are fuzzy concepts. The region north of a point can be 
bounded by two vectors pointing NE and NW (the trian- 
gular model), but this model fails when computing north 
of an object that is elongated in the E-W direction. Some 
spatial relations like "in front of, .... behind," and "be- 
yond" are relative to the observer's direction. Although 
these are not very frequent (only 25 cases in the 1340 
sample descriptions), they still appear. Other spatial re- 
lations like "above" are dependent on understanding the 
slope of elevation around the object. 
To solve some of these problems, we have developed 
and implemented the LEI system to partially automate 
interpretation of this sublanguage. LEI is described in 
the following section. 
4 The LEI System 
4.1 Organization 
The LEI 3 (Location and Elevation Interpreter) system is 
an implementation of our algorithms for interpreting the 
sublanguage of location description labels for biological 
specimens. LEI is composed of four main components: 
the language analyzer PPI, the geographical reasoner GR, 
the user interface LEIview, and the geographic knowl- 
edge base GKB. The geographic knowledge base con- 
tains an object-oriented description of geographical ob- 
jects such as valleys, streams, and waterfalls with their 
associated locations and names. The user interface dis- 
plays maps and allows users to add or modify object lo- 
cations. The language analyzer parses the English loca- 
tion description and produces a collection of spatial rela- 
tions that relate the actual collection point to geograph- 
ical objects. It uses knowledge of geographical objects 
3Leiis also the Hawaiian word for "garland," typically made 
out of flowers, leaves, or feathers. 
139 
and their associated names from the geographic knowl- 
edge base. The geographical reasoner translates spatial 
relations from the language analyzer into polygons and 
performs polygon intersection calculations to obtain the 
area specified by the spatial relations. Each component 
is described in detail below. 
4.2 GKB, the Geographic Knowledge Base 
LEI uses three U.S. Geological Survey (USGS) digital 
cartographic databases as the starting point for GKB, the 
Geographic Knowledge Base. These include the DLG 
(Digital Line Graph), GNIS (Geographic Name Informa- 
tion System), and DEM (Digital Elevation Model). Un- 
fortunately, these databases are not object-oriented, that 
is, they do not link the names in GNIS to the object loca- 
tions in DLG. The GNIS database contains only names, 
USGS quadrangles, a feature class, and the coordinates 
of the name as it appears on a USGS map. The DLG 
database contains a hierarchical organization of points, 
line segments (composed of points), and areas (composed 
of line segments) along with a two-level type hierarchy 
composed of major and minor codes. Unfortunately, 
the 60 plus GNIS feature classes do not correspond to 
the over 200 DLG major and minor codes. The DEM 
database consists of a raster style set of elevation values. 
To convert the three USGS databases into a useful 
object-oriented format requires human intervention to as- 
sociate the names with the line and area objects (point 
objects are already correctly located by the GNIS coordi- 
nates). This process uses the LEIview component to dis- 
play the appropriate section of the map to the user, then 
displays names one by one along with the geographically 
closest object of the same type as the name. The user can 
confirm the match, ask for alternatives, or modify the set 
of line segments or areas to give the actual location of the 
named object. In many cases, there are no corresponding 
objects, so the user must draw the outline of the area from 
scratch. This is required for all valleys and mountains 
since these are missing from the DLG specifications. 
Determining the closest object of the same type re- 
quires matching the GNIS feature class to appropriate 
DLG major and minor codes. This is done using LEI's 
own type hierarchy that includes type classes correspond- 
ing to each GNIS feature class and to each DLG major 
and minor code along with many bridging type classes 
and higher level types. Given a GNIS feature class, LEI 
first indexes into LEI's type hierarchy to find the cor- 
responding LEI type. If this type has a corresponding 
DLG code, then that is the most likely match. Less likely 
matches consist of any subtypes that might have associ- 
ated DLG codes. If there are no DLG codes at this type 
level nor at subtype levels, then LEI searches up the hi- 
erarchy for supertypes that have associated DLG codes. 
Using this algorithm, the matching process manages to 
find the correct match most of the time, so the user's time 
is freed to worry about the many missing entries and er- 
rors in USGS databases (e.g., rivers that extend into what 
should be coastlines, disconnected lines, etc.). 
4.3 LEIview, the User Interface 
The LEIview component provides a graphical interface 
that allows users to view maps; zoom; scroll; rearrange, 
add, and delete layers of the map (including DLG lines, 
GNIS names, and DEM elevations); search for named ob- 
jects; enter points, line segments, or areas for new ob- 
jects; modify existing objects; and view the results of 
interpreting location descriptions (both the English de- 
scription and the area resulting from processing are dis- 
played). LEIview is written in C under X windows with 
Motif widgets. 
LEIview is used to associate names with object loca- 
tions in building the GKB geographic knowledge base. It 
is also used to display the results of interpreting location 
descriptions. When there are sections of the description 
that are not comprehensible to the PPI language analyzer, 
LEI sends the description to LEIview, which displays the 
description with the incomprehensible parts highlighted 
and displays the regions corresponding to the understood 
portions of the description. The user can tell LEI to ig- 
nore the unknown parts of the description, delay process- 
ing this description until later, send the description back 
for reprocessing, or add new geographic objects by en- 
tering new points, line segments, or areas and selecting 
the corresponding words in the description. Any new ob- 
jects are stored in GKB and the correspondence between 
the words and the new object are stored in the PPI lan- 
guage knowledge base. 
4.4 PPI, the Language Analyzer 
The PPI (Prepositional Phrase Interpreter) component is 
responsible for parsing the natural language location de- 
scriptions and converting them into spatial relations. PPI 
uses the PAU 4 parser and understander \[Chin, 1992\] to 
interpret the English descriptions and convert them into 
spatial relations represented in the MERA (Meta En- 
tity Relation Attribute) semantic-network-style knowl- 
edge representation language \[Takeda et al., 1992\]. PAU 
is an all-paths, chart-based, unification parser that com- 
pletely integrates syntactic and semantic processing. 
Figure 1 shows the MERA graph for the grammar rule, 
PP ~-- Prep NP (i.e., a Prepositional-Phrase is a Preposi- 
tion followed by Noun-Phrase), along with its semantic 
interpretation. The node PP-pat represents the left-hand- 
side of the rule, and the relations Pea (pattern component 
A) and Pcb (pattern component B) point to the compo- 
nents on the right-hand-side of the rule. The Ref relation 
denotes the meaning of the rule: a Geographic-object that 
has a Spatial-relation to some other Geographic-object. 
The Unify relation between the Prep and the Spatial- 
relation indicates that the meaning of the Prep should be 
unified with the relation, Spatial-relation. Likewise, the 
Unify relation between the NP and the lower Geographic- 
object indicates that the meaning of the NP should be uni- 
fied with the lower Geographic-object Figure 2 speci- 
fies the meaning of the Prep, "along" as an instance of 
4pau is also the Hawaiian word for "finished." 
140 
PP-pat Ref ~ Geographic-object 
\[Spatial-relation 
Prep N .~nify \[ 
Unify-""---,~ Geogtraphic_object 
Figure 1: Rule for PP ~-- Prep NP. 
PP-pat Ref ~ Geographic-object 
preP~~Np /Near 
Figure 3: The PP-pat rule after parsing "along" and be- 
fore parsing "Ainapo." 
the Spatial-relation Near relating a Geographic-object to 
a Linear-object (a subtype of Geographic-object). 
Geographic-object 
Prep 
word=along 
Ref Near 
Linear-object 
Figure 2: The meaning of the preposition "along." 
The interpretation of the PP "along Ainapo" demon- 
strates how the integration of syntactic and semantic pro- 
cessing in PAU allows the early use of semantic con- 
straints to reject semantically anomalous parses. "Along 
Ainapo" is ambiguous because Ainapo is both a trail and 
an area. However, since "along" only applies to linear 
objects such as trails, the Ainapo area interpretation is 
rejected by PAU. This happens when PAU is applying 
the grammar rule of Figure 1. When unifying the mean- 
ing of the Prep "along" (shown in Figure 2) with the 
Spatial-relation, the result is a Near relation. However, 
the sources and sinks of both relations must also be uni- 
fied. This changes the lower Geographic-object into a 
Linear-object as seen in Figure 3, which shows the state 
of the "PP ~-- Prep NP" rule just before parsing "Ainapo." 
In PAU, both meanings of Ainapo are tried in parallel. 
The area meaning of Ainapo is rejected because an Area- 
object cannot unify with a Linear-object. This leaves 
only the Ainapo trail meaning to parse successfully. 
Table 1 shows the spatial relations in PPI along with 
the corresponding prepositions. 
4.5 GR, the Geographical Reasoner 
The GR (Geographical Reasoner) component takes the 
output from the PPI component, which is a set of spatial 
141 
relations, converts these into polygons, performs poly- 
gon overlay operations to find the common intersection 
of all the polygons, computes the center of the mini- 
mum bounding rectangle (mbr) of the polygon intersec- 
tion, then returns the coordinates and centroid of the mbr. 
GR like PPI is written in Common LISP and converses 
with LEIview through UNIX sockets. 
The first step is the most difficult since there are no 
generally accepted algorithms for converting spatial rela- 
tions into areas. For the spatial relations based on cardi- 
nal directions such as East-of, Peuquet and Zhan (1987) 
give a complex algorithm for determining if one poly- 
gon is in a particular directional relationship with another. 
Their algorithm takes into account the shapes of the poly- 
gons (e.g., east of an elongated N-S polygon covers a dif- 
ferent area than east of a small point polygon) and con- 
siders polygons that partially enclose or intertwine one 
another. Their algorithm is a refinement of the basic trian- 
gular model (in which North is the open-ended triangular 
region between two vectors pointing NE and NW), but 
it still does not give any limits concerning the distance 
between the polygons. Unfortunately there is no abso- 
lute distance that forms the edge of the region North-of 
some polygon. In a sense, the edge is given by the limit 
of human sight in that direction. 
The algorithm currently used in GR for interpreting 
cardinal-direction relations around an geographic object 
starts by computing the minimum bounding rectangle 
(mbr) for the object. The area next to the mbr with the 
same size as the mbr is taken as the meaning of the spa- 
tial relation. Since the resultant area is the same size as 
the original object, this makes the meaning of cardinal di- 
rections relative to the size of the reference object, taking 
into account the fact that larger objects are visible from 
farther away. Cardinal directions relative to point objects 
are interpreted as a square, 500 meters on a side, lying in 
the appropriate direction. 
The observer or object oriented relations such as 
Adjacent-to ("beside waterfall," "on Kona-Hilo Hwy"), 
Beyond ("1 1/2 mile beyond end of 20 Mile Road," "at 
back of Waihoi Valley"), Front-of (no examples in the 
sample data), Right-of ("right hand side of Kupu Kai 
Gap"), and Left-of ("Kulani Prison Road, toward Ku- 
lani Prison from intersection w/Volcano Road, left road- 
Spatial-relation Prepositions 
Adjacent-to adjacent to, beside, next to, on 
At-elevation above, at, below, down, up 
Between between 
Beyond \[in/at\] back of, behind, beyond 
East-of east of 
From from 
Front-of before, \[in\] front of 
Left-of \[to \[the\]\] left \[hand side\] of 
Near 
North-of 
Right-of 
adjacent to, along, around, at, by, near, outside \[of\] 
north of 
\[to \[the\]\] right \[hand side\] of 
South-of south of 
Toward to, toward 
West-of west of 
Within among, at, in, inside, into, on, on top of 
Table 1: Spatial Relations and Prepositions in PPI. 
side") require understanding the orientation of the object 
or observer. Currently in GR, only object-oriented rela- 
tions are processed. Given an object with a front, back, 
and sides (left and right), the corresponding relations are 
Front-of, Beyond, and Adjacent-to (Left-of and Right- 
of). These areas are calculated from the object's mbr in a 
similar fashion to the cardinal direction relations. 
The At-elevation relation with respect to a given al- 
titude requires computing the subregion of the common 
intersection area that is within 40 meters 5 of the given 
elevation. The At-elevation relation relative to an object 
(e.g., "above Schofield") requires computing the preva- 
lent slope of the terrain around the object. GR takes a 
200 meter square on the up/down side of the object. 
Between, From, and Toward are handled by taking the 
mbr of the two objects, then computing the two corner 
points on each mbr that is furthest on either side from 
the line connecting the centers of the mbrs. These four 
points are then connected to form the area between the 
two objects. 
The Near relation is converted into a buffer zone 
around the area. Currently GR uses a fixed distance 
of 200 meters for simplicity, however further study is 
needed to determine if this corresponds to most people's 
interpretation. There may be individual, cultural, or re- 
gional differences in interpretation. Also, the size of the 
buffer zone may depend on the size of the geographic ob- 
ject. 
Currently GR does not handle references to terrain 
type, wetness, or typical sun exposure because this type 
of data is not available in the USGS databases. Refer- 
ences to ordinal forks and branches are assumed to start 
from the head of the rivers or valleys. Generic terms are 
handled after processing all other spatial relations by ex- 
haustively searching for any instances of the same type 
5USGS DEM data have a vertical resolution of one meter 
and a horizontal resolution of thirty meters. 
142 
(or subtypes) that intersect with the intersection of the 
other known areas. In cases of multiple matches, the user 
is asked to help disambiguate through LEIview. 
5 Future Directions 
Because collectors often collect specimens on trips (ei- 
ther day hikes or multi-day camping expeditions), an 
analysis of the path of the collectors should yield valu- 
able information about the location of collection. Speci- 
mens are typically labeled with the collection date and the 
collector's accession number, which provides the relative 
time of collection for specimens on that day. This infor- 
mation can be used to disambiguate location descriptions 
and to pinpoint vague locations. For example, in Hawaii, 
there are not only two Waihee Streams, but also a Wai- 
hee River. In a location description that mentions, "along 
Waihee Stream," there is ambiguity as to which of these 
three waterways is actually meant. In the current version 
of LEI, disambiguation is possible only if the description 
contains more information that specifies an area that in- 
tersects with only one of the three streams. By adding 
reasoning about time using accession dates and numbers 
and combining this with reasoning about paths, LEI could 
determine that it is unlikely that the collector stopped col- 
lecting specimens along one Waihee Stream, flew to an- 
other island to collect a specimen along another Waihee 
Stream, then flew back to continue collecting along the 
first Waihee Stream. 
This type of reasoning can also help to pinpoint 
which part of Waihee Stream is meant by "along Wai- 
hee Stream." If LEI knows that the previous specimen 
was collected at point A and the following specimen was 
collected at point B, then LEI can make the reasonable 
assumption that this specimen was collected somewhere 
near the intersection of Waihee Stream and a region be- 
tween points A and B. Using this type of reasoning, LEI 
can even make a reasonable guess about the collection 
location of specimens that have no location labels (pro- 
vided only that they have an accession number and acces- 
sion date given by the collector and the adjacent specimen 
numbers can be located). Adding such reasoning about 
time and paths would improve the accuracy of LEI's pro- 
cessing. 
6 Conclusions 
The LEI system demonstrates the feasibility of under- 
standing the sublanguage used in location descriptions 
for biological specimens. Although this is an important 
and valuable task in and of itself, there is a much greater 
potential for application of the NLP and geographical 
reasoning techniques demonstrated in LEI to other areas 
such as natural language interfaces to general GISs (Geo- 
graphic Information Systems). There is a need for valida- 
tion of these techniques and a study is currently planned 
to compare the results of LEI with results obtained man- 
ually. Finally, the problems encountered in building LEI 
point to several new directions. First, the GKB com- 
ponent shows how object-oriented geographic databases 
should be organized in the future. Second, many new 
studies are required to determine the limits of fuzzy spa- 
tial relations like North-of, Front-of, and Near. Such 
studies should investigate task dependencies, context de- 
pendencies, individual variances, and cultural/regional 
variances. Such studies would lead to advances in under- 
standing human cognition of spatial relations that would 
be directly applicable in GISs like LEI. 
7 Acknowledgements 
This research was sponsored in part by the National Sci- 
ence Foundation B S R&SES Grant No. B SR-9019041. 

References

Kuipers, B. J. (1985). Modeling Human Knowledge 
of Routes: Partial Knowledge and Individual Variation. In the Proceedings of the Third National Conference on Artificial Intelligence, pp. 216--219. 

Mark, D.M. and A.U. Frank (1991). (Eds.), Cognitive and Linguistic Aspects of Geographic Space, 
Klewer Academic Publishers, Boston. 

Peuquet, D. and Z. Ci-Xiang (1987). An Algorithm 
to Determine the Directional Relationship between 
Arbitrarily-Shaped Polygons in the Plane. In Pat- 
tern Recognition 20(1), pp. 65-74. 

Takeda K., D. N. Chin, and I. Miyamoto (1992). 
MERA: Meta Language for Software Engineering. 
In the Proceedings of the 4th International Conference on Software Engineering and Knowledge Engineering, Capri, Italy, June, pp. 495-502. 

Talmy, L. (1983). How Language Structures Space. In 
H. Pick and L. Acredolo, Eds., Spatial Orientation: 
Theory, Research, and Application Plenum Press, 
New York, pp. 225-282. 

Andre, E., G. Bosch, G. Herzog, and T. Rist (1986). 
Coping with the Intrinsic and Deictic Uses of Spatial 
Prepositions. In Ph. Jorrand and V. Sgureg (Eds.), 
Artificial Intelligence II, Proceedings of AIMSA-86, 
pp. 375-382. 

Chin, D. N. (1992). "PAU: Parsing and Understanding 
with Uniform Syntactic, Semantic, and Idiomatic 
Representations." In Computational Intelligence, 
8(3), pp. 456-476. 

Davis, E. (1986). Representing and Acquiring Geo- 
graphic Knowledge. Morgan Kaufman, Los Altos, 
CA. 

Frank, A. (1991). Qualitative Reasoning about Cardinal 
Directions. In D. Mark and D. White (Eds), Pro- 
ceedings of Autocarto 10, pp. 148-167. 

Freeman, J. (1975). The Modeling of Spatial Relations. 
In Computer Graphics and Image Processing, 4, pp. 
156--171. 

Herskovits, A. (1986). Language and Spatial Cogni- 
tion. Cambridge University Press, Cambridge. 
