snodgrass (1980) a standardized set of 260 pictures. norms for name agreement, image agreement, familiarity, and visual complexity

Journal of Experimental Psychology:Human Learning and Memory1980, Vol. 6, No. 2, 174-215

A Standardized Set of 260 Pictures: Norms forName Agreement, Image Agreement, Familiarity,

and Visual Complexity

Joan Gay Snodgrass and Mary VanderwartNew York University

In this article we present a standardized set of 260 pictures for use in ex-periments investigating differences and similarities in the processing of pic-tures and words. The pictures are black-and-white line drawings executedaccording to a set of rules that provide consistency of pictorial representation.The pictures have been standardized on four variables of central relevance tomemory and cognitive processing: name agreement, image agreement, fa-miliarity, and visual complexity. The intercorrelations among the four mea-sures were low, suggesting that the)' are indices of different attributes of thepictures. The concepts were selected to provide exemplars from several widelystudied semantic categories. Sources of naming variance, and mean familiarityand complexity of the exemplars, differed significantly across the set of cate-gories investigated. The potential significance of each of the normative vari-ables to a number of semantic and episodic memory tasks is discussed.

Investigators studying aspects of verbalprocesses have long had access to extensivenormative data on various objective andsubjective dimensions of their verbal mate-rials. Brown (1976) recently compiled acatalog of scaled verbal materials thatincluded 172 studies providing such infor-mation. For the set of verbal materialsmost comparable to the present set ofpictures—English nouns—such dimensionsinclude objective measures of frequency ofoccurrence and subjective measures offamiliarity, age of acquisition, concreteness,imagery, meaningfulness, and emotionality.

In contrast, normative data on character-istics of pictorial representations of concrete

This research was supported by Research GrantMH-20723 from the National Institute of MentalHealth, U.S. Public Health Service, and by a Bio-medical Research Support Grant to New YorkUniversity from the National Institutes of Health,U.S. Public Health Service.

The authors would like to thank MaryannMcManus, Anthony Vallone, Spencer Fischer, andCarolyn Gleeman for help at all stages of thedesign, data collection, and analysis.

Requests for reprints should be sent to JoanGay Snodgrass, Department of Psychology, NewYork University, 6 Washington Place, New York,New York 10003.

nouns, whose use in experiments in bothsemantic and episodic memory tasks hasproliferated in recent years, have beenlacking. Because there are so many differentways to draw even the simplest object, eachinvestigator has been forced to develop hisor her own set of pictures that mustnecessarily be highly idiosyncratic. Wecannot assume that the results of studiesemploying such different representationsof the same concepts are comparable. Inaddition, the degree to which each picturepossesses characteristics that affect theprocess under investigation is unknown.

In recognition of this long overdue needfor a standardized set of picture stimuli,we present here a new set of 260 picturesthat are standardized on several variablesof potential importance to cognitive proces-sing tasks. The pictures are presented inAppendix A and the accompanying normsin Appendix B. In addition, slides of thepictures and their names are availablefrom a commercial publisher.1

1 Slides of the pictures or of their names maybe ordered from Life Science Associates, 1 FenimoreRoad, Bayport, NY 1170S.

Copyright 1980 by the American Psychological Association, Inc. 0096-1515/80/0602-0174$00.7S

174

NORMS FOR PICTURE STIMULI 175

In the next sections, we first brieflyreview evidence for processing differencesbetween pictures and words in semanticand episodic memory tasks, and thendescribe the assumptions and procedureswe used to construct the picture set.

Processing Differences BetweenPictures and Words

Processing of pictures and words hasbeen compared in many different tasksand for many different reasons. We canclassify these tasks into two broad categor-ies—those tapping semantic memory, ageneral knowledge store shared by mostsubjects, and those tapping episodic mem-ory, an event knowledge store about theoccurrence of events in time and spacethat is idiosyncratic to individual subjects(Tulving, 1972). Semantic memory tasksinclude naming pictures and words, gen-erating images to words, categorizingpictures or their names, and makingsymbolic comparison judgments aboutwhether X has more of some propertythan Y. Episodic memory tasks includethe standard verbal learning and memoryparadigms, such as free and serial recall,paired-associate recall, yes/no or forced-choice recognition, serial and spatial recon-struction and recognition, and so forth.The questions that emerge from these twoclasses of tasks are, first, how pictures andwords differ in their access to semanticmemory, and second, how they differ intheir registration in episodic memory. Inthis necessarily brief review of empiricaldifferences between pictures and wordswe shall consider two semantic memorytasks—naming and categorization—andtwo episodic memory tasks—recognitionand free recall.

Semantic Memory Tasks

Naming. Naming latencies, as measuredby voice key activation, are faster forwords than for pictures (Cattell, 1886;Fraisse, 1960; Potter & Faulconer, 1975).Furthermore, certain characteristics of thepicture's name, such as frequency in print,

have much larger effects on naminglatencies of pictures than on readinglatencies of the picture's name (Oldfield &Wingfield, 1965). However, there is dis-agreement about which of several variablescorrelated with frequency in print isprimarily responsible for the differences inpicture-naming latencies. Carroll and White(1973b) found that age of acquisition of aconcept's name was more predictive ofnaming latencies than frequency and, infact, that frequency was insignificant inpredicting naming latencies when theage-of-acquisition variable was included inthe multiple regression. Lachman (1973)measured the degree to which subjectsagreed on the name of the picture withthe information statistic H and found thatH was more predictive of naming latenciesof pictures than either age of acquisitionor frequency in print, although the lattertwo variables generally predicted significantamounts of variance in the naming laten-cies. However, individual subjects variedin the degree to which each of the threevariables predicted latencies. To under-stand the picture-naming process, then,it is clear that one must know at least thefollowing characteristics of each picture—the frequency of its name in print, the ageof acquisition of its name, and its degreeof name agreement.

Categorization. Tasks in which themeaning of pictures or their names mustbe accessed have been used to study thequestion of whether pictures and theirnames access a common semantic represen-tation. In a categorization task, there isample evidence that pictures are categor-ized faster than their names. For example,in a yes/no task in which the category isgiven ahead of time, pictures are categor-ized either faster than words (Potter &Faulconer, 1975) or with the same speed(Smith & Magee, in press). In a same-different task, when two instances of pic-tures or words in a same or different cate-gory are presented simultaneously, picturesresult in faster match or mismatch decisionsthan words (Pellegrino, Rosinski, Chiesi,& Siegel, 1977). Furthermore, comparisonsbetween picture and word forms of exem-

176 JOAN GAY SNODGRASS AND MARY VANDERWART

plars of categories were made with latenciesintermediate to the picture-picture andword-word comparisons. These data maybe interpreted as indicating that picturesand words access a common semanticrepresentation that is neither verbal norimaginal in nature. That is, if determiningmembership were dependent on accessingthe exemplar's name, then pictures wouldtake longer to categorize than namesbecause naming pictures takes longer thannaming words.

However, these data could as easily beaccounted for by assuming that categoriza-tion is based on access to imaginal codes.This would account for faster categorizationof pictures than their names and for thedata of Pellegrino et al. (1977) on mixedform comparisons. This hypothesis is alsoconsistent with Paivio's (1978) recentargument that such attributes as size,value, and intelligence are stored at theimage level and may be in the form ofremembered motor/visceral reactions.

Episodic Memory Tasks

In episodic memory tasks, an impressivearray of evidence suggests that pictorialmaterial is remembered better than verbalmaterial, when memory is tested either byrecognition (Nickerson, 1965; Shepard,1967; Snodgrass & McClure, 1975; Stand-ing, Conezio, & Haber, 1970) or by recall(Bousfield, Esterson, & Whitmarsh, 1957;Paivio, Rogers, & Smythe, 1968). Thereare at least three hypotheses that havebeen proposed for the superiority ofpictures over words: dual coding, superiorsensory codes, and uniqueness of entry insemantic memory.

According to Paivio's dual coding model,pictures are better remembered than wordsin most tasks for one or both of two reasons:First, pictures are more likely to be duallycoded than words (i.e., registered in boththe image and verbal stores); and second,the image code, which is more likely to bestored to a picture than to a word, is themore effective code for item memory(Paivio, 1971; Paivio & Csapo, 1973).For example, when subjects are instructed

to form visual images of words and toname pictures (which presumably providesdual codes for both types of materials),recall performance is equal for pictures andwords (Paivio & Csapo, 1973), althoughrecognition performance continues to besuperior for pictures over words under suchdual-coding instructions (Snodgrass &McClure, 1975). Further evidence thatpictures are more likely to be duallyencoded than words, in the absence ofinstructions, comes from the finding thatsubjects have difficulty deciding betweena studied picture and its name, in aforced-choice recognition test (Snodgrass,Wasser, Finkelstein, & Goldberg, 1974),and from the lack of improvement in itemrecognition memory for pictures studiedunder verbal encoding instructions overthose studied under imagery instructions(Snodgrass & McClure, 1975). In contrastto this evidence, Intraub (1979) has foundthat picture-naming latency and presenta-tion time do not interact in a recognitionor recall memory task, thereby providingno evidence that implicit naming improvedpicture memory at slow presentation rates.

The superior sensory code hypothesis,proposed by Nelson and his colleagues,attributes the pictorial superiority effectto the more elaborate sensory codes ofpictures as compared to words (Nelson,Reed, & McEvoy, 1977; Nelson, Reed, &Walling, 1976). They found that increasingthe visual similarity of a set of pictures ina paired-associate recall task, while keepingconceptual similarity constant, had theeffect of destroying, and even reversing,the superiority of pictures over words.

The third hypothesis, that pictures areremembered better than words becauseof their uniqueness in semantic memory,has been proposed by Durso and Johnson(1979) and Snodgrass (Note 1). Theargument here is that words, even concretenames, are more polysemous than pictures,and hence their semantic representation isless likely to be contacted on a recognitiontest or retrieved during recall than theword's corresponding picture. This argu-ment is similar to that made by critics of theencoding specificity hypothesis of Tulving


and Thomson (1971), who argue thatchanging word context from study to testeffectively changes which entry is accessedin semantic memory (Anderson & Bower,1974; Martin, 1975; Reder, Anderson, &Bjork, 1974). A related hypothesis has beenproposed by Potter, Valian, and Faulconer(1977) to account for the finding thatalthough relatedness of a probe to asentence could be decided as quickly forpicture as for word probes, suggesting anabstract format for sentence meaning,probe pictures were recalled better thanprobe words. They attribute the superior-ity of pictures to greater specificity ofmeaning of the picture.

Of course, the three hypotheses are notmutually exclusive: It could be the casethat all three attributes of pictures—theirgreater probability of dual coding, theirmore elaborate sensory codes, and theiruniqueness of meaning—jointly accountfor the superiority of pictures over wordsin episodic memory.

Regardless of the theoretical argumentsadvanced for the differential efficiency ofpictures and words in contacting informa-tion in semantic memory or of beingregistered in episodic memory, it is clearthat there are important parameters char-acterizing the interlingua between picturesand their names that need to be established.It is the aim of the present study toinvestigate and measure some of theseparameters.

Development of the Present StudyThe present study evolved in two steps.

First, we selected the concepts to be drawnand established a set of guidelines for themanner in which they were to be drawn.Second, we decided which of severalpossible types of information about thepictures we would collect.

The 260 concepts were selected on thebasis of three criteria: first, that they beunambiguously picturable; second, thatthey include exemplars from the widelyused category norms of Battig and Monta-gue (1969); and third, that they representconcepts at the basic level of categorization(Rosch, Mervis, Gray, Johnson, & Boyes-Braem, 1976).

The first criterion, that concepts beunambiguously picturable, is a matter ofdegree. What we mean by this, opera-tionally, is the degree to which subjectswill show consensus about the name togive the object. There are several sourcesof naming ambiguity for pictures. Thesesources stem from the concept itself, thepicture itself, the possible name or namesa picture can have, and the subjects'knowledge of concepts and their names.

Some concepts are, by their very nature,not unambiguously picturable. These in-clude concepts of relationship (father,sister), mass nouns (money, water, grass),and abstract concepts (liberty, time,honesty). Some pictures, by the way theyare drawn, are not easily recognizable forwhat they are. That is, they are drawn inan atypical way or with too little detailto make their identity clear. As we shalldiscuss, we have attempted to avoid thisproblem. Some concepts simply can becalled by more than one name, even thoughthe names appear to mean the same thingto subjects. Object names that appear tobe nearly synonymous include phonograph/record player, trumpet/horn, and purse/pocketbook. We have included such con-cepts in the set and have indicated thecases in which synonymous terms accountfor name ambiguity. Finally, some concepts,because they are unfamiliar to subjects,also are not known to them by name,even though they are recognizable asexemplars of a particular category. Forexample, artichoke and chisel may not beknown by name, but subjects are awarethat they belong to the categories ofvegetables and tools, respectively. Thesesources of ambiguity are also identified inthe norms.

The second criterion, that the set ofpictures include exemplars from well-studied categories, is important because oftheir potential use in category-judgmenttasks. We have used the Battig andMontague (1969) category norms as aguide. Of the 56 categories studied byBattig and Montague, 15 contain many

(text continued on p. ISO)


Table 1Concepts and their Ranks, from the Battig and Montague Category Norms

Concept

Four-footed animal

AlligatorBearCamelCatCowDeerDogDonkeyElephantFoxFrogGiraffeGoatGorillaHorseKangarooLeopardLionMonkeyMousePigRabbitRaccoonRhinocerosSheepSkunkSquirrelTigerTurtleZebra

Rank

(30)

639

2824

121

19.57

23751415863

7521.55

32.5108

19.532.5271336176

40.516

Kitchen utensil (14)

BowlBroom

*ClockCupForkFrying panGlassKettleKnifePotRefrigeratorRolling pin

^ScissorsSpoonStove

"TableToaster

Article of furniture

AshtrayBedChair

99876113

152144.5

15

18.528.57628

49.522

(14)

49.531

Concept Rank

Article of furniture (continued)

ClockCouchDeskDresserLamp

* PianoRecord player

* RefrigeratorRocking chairStool

*StoveTableTelevisionVase

647586

17553049.51035

29

64

Part of the human body (12)

ArmEarEyeFingerFootHairHand

fHeartLegLipsNoseThumbToe

Fruit (11)

AppleBananaCherryGrapesLemonOrangePeachPearPineappleStrawberry

"TomatoWatermelon

Weapon (7)

*Airplane*ArmAxe

*Bat (baseball)*BookBottleCannon

*Car

28475

139

151

33.56

4510

1476

10253

13161517

39.51091653

10961.51030.5


Table 1 (continued)

Concept Rank

Weapon (continued)

Chain 22"Chair 87.5*Foot 44.5*Fork 87.5Gun 2

"Hammer 24*Hand 26

t*Knife 1*Nail 109Scissors 44.5

"Screwdriver 87.5"Shoe 87.5"Wrench 69

Carpenter's tool (12)

"Axe 28t*Box 73

Chisel 7Hammer 1

t*Knife 30Ladder 31.5Nail 3Nut 34Pencil 14.5Pliers 10Ruler 8Saw 2

*Scissors 51.5Screw 12Screwdriver 4Wrench 9

Article of clothing (19)

Belt 16Blouse 5.5Boot 40Cap 38Coat 7Dress 8Glove 15Hat 9Jacket 13.5Pants 3Ring 40Shirt 1Shoe 4Skirt 5.5Sock 2Sweater 10Tie 11Vest 26Watch 46.5

Part of a building (3)

"Chair 41.5"Desk 52

Concept Rank

Part of a building (continued)

DoorDoorknob

fLock"NailWindow

295

123521

Musical instrument (9)

Accordiont*Bell

DrumFluteFrench hornGuitflrHarpPianoTrumpetViolin

Bird (8)

ChickenDuckEagleOstrichOwlPeacockPenguinRooster

Type of vehicle

Airplanet*Balloon

BicycleBusf*CarHelicopter

"HorseMotorcycleRoller skateSledTrainTruck

t*Wagon

Toy (18)

"AirplaneBaby carriageBallj-Fdit

BalloonBat (baseball)

t*BearBell

"BicycleBook

"Boot

2434

26

177

12134

2123.5

533.523.53740.553.5

(10)

353.562j

2632

7373445

10

13.554.5

2269

9376.52235

123


Table 1 (continued)

Concept Rank

Toy (continued)

Box*CannonTor\*al

*Cat*ClockClown

*DogDoll

*Drumt*Duck

Footballt*Glove*Gun"Hammer*Hat*HorsefHouseKite

*Knife* Roller skateSailboat

*Sled*StoveSwingTelephoneTop

TrainTruck

93123

393

12376.5511

351233545.54.5

519330.55763.563.551

12363.576.545.5931064.5

Concept Rank

Toy (continued)

WagonWhistle

Vegetable (13)

ArtichokeAsparagusCarrotCeleryCornLettuceMushroomOnionPeanut

*PearPepperPotatoPumpkinTomato

"Watermelon

Insect (8)

AntBeeBeetleButterflyCaterpillarFlyGrasshopperSpider

1563.5

3291

1137

46183646265

466

46

23.56

1421195

Note. A dagger indicates that the picture does not represent the meaning intended by the name, and anasterisk indicates that this exemplar appears with higher frequency in another category in the table (exceptwhen it is daggered). Frequencies in parentheses beside each category indicate the number of unasteriskedor undaggered concepts.

picturable exemplars, such as four-footedanimals, kitchen utensils, and human bodyparts. Table 1 lists the 189 conceptsincluded in the set of pictures that aremembers of one or more of these 15Battig and Montague categories. Onlycategories having more than five exemplarsin the picture set are included, and onlynames given by more than one respondentin the Battig and Montague study areshown in the table.

The rank by frequency of the concept ineach category is shown next to its name.Concepts with high ranks, such as dog forfour-footed animal and chair for articleof furniture, were given by many subjects,

whereas those with low ranks, such asgorilla for four-footed animal and vasefor article of furniture, were given by fewsubjects. Thus it is possible to vary thedegree to which category exemplars areprototypical of the category.

A number of the concepts appear inmore than one category. For example,airplane appears as a weapon, a type ofvehicle, and a toy. Of course, multiple-category concepts differ in their rankingacross the categories; those concepts thatare of lower typicality (e.g., airplane as aweapon and toy) are identified withasterisks. Furthermore, the particular senseof a name is not always instantiated in its


picture. For example, the knife, which islisted as the most typical exemplar of thecategory weapon, is surely not the sameknife that is listed as the most typicalexemplar of the category kitchen utensil inits imaged or pictured form. The picture ofa knife included in the present set corre-sponds to the second sense rather than thefirst. Similarly, the pictured heart does notrepresent the sense of the heart given as anexemplar of the category part of the humanbody. Pictures that in our judgment do notrepresent the meaning of the categoryexemplar are indicated by a dagger inTable 1. Thus, the rank of the Battig andMontague names will correspond to therank of their corresponding pictures onlyto the extent that the picture representsthe meaning intended in the name givenby respondents.

The 71 remaining pictures, not exemplarsof any of the Battig and Montague cat-egories listed in Table 1, are not assignableto any well-defined category (e.g., anchor,barrel, flag), or are members of categoriesfrom the Battig and Montague normshaving too few exemplars in this pictureset to be useful (e.g., house, church, andbarn), or are members of small categoriesnot studied by Battig and Montague (e.g.,cigar, cigarette, and pipe; needle, thread,and thimble). These concepts were includedin the set either because they are easilydrawn or because they have been used byother investigators.

The third criterion, that the conceptsbe represented at their basic level, seemsto have been implicitly followed by mostinvestigators using pictures as stimuli.For example, pictures of the categoriesanimal, tool, and vegetable are not used(indeed, it is not clear whether any pictorialrepresentation could elicit these names).But, pictures of the categories bird, tree,flower, and fish, some of which Roschet al. (1976) have shown are basic cat-egories, have been used. Indeed, suchatypical exemplars of the category bird aschicken, duck, and ostrich also appear tobe basic level categories. We suggest, infact, that another criterion for identifying

whether a concept is at a basic level iswhether its picture produces consensus innaming.

Once the concepts were selected, someguidelines had to be established for howthey should be drawn.2 The criteria weused were that the drawing be correct(i.e., realistic) in its details, that thedrawing be the most typical representationof a concept, that the drawing be anunambiguous representation of the concept,and that the drawing include the amount ofdetail needed to be consistent with thecomplexity of the real-life object. Althoughthese criteria are admittedly subjective, apanel of four judges (consisting of theauthors and two students) reached a highdegree of consensus about when a drawingdid, or did not, fulfill the criteria.

Another decision concerned the orienta-tion of each figure. We adopted the follow-ing rules: (a) Animals are shown insideways view, with approximately equalnumbers facing left and right; (b) objectswhose up-down orientation may vary (e.g.,fork, chisel) are drawn with the functionalend down; (c) long, thin objects areoriented at a 45° angle, with approximatelyequal numbers in the two possible orien-tations.

The objects were drawn so as to be ofapproximately equal subjective size basedon judgments of the two authors. Toaccomplish this, it was necessary to drawobjectively smaller objects as slightlysmaller in their pictured form than objec-tively larger objects. Because these subjec-tive size decisions are very difficult to make,the size adjustments were made within cat-egories. Thus, the drawing of mouse issmaller than that of lion, and the drawingof strawberry, smaller than that of apple.

We collected four measures on the set of260 pictured concepts, based on the four

2 The pictures were drawn by the graphic designfirm A Good Thing, Inc., 230 Park Avenue, NewYork, NY 10017.


cognitive processes of naming, imaging,familiarity judgments, and complexityjudgments. The four measures seemed tous to represent distinctly different aspectsof pictured stimuli. First, we had subjectsname the pictures to determine the picture'smost common name and the degree towhich subjects agreed on the name. Thisinformation is of obvious importance innaming latency studies, picture-namematching studies, recall memory studies,and recognition studies in which verbalencoding is manipulated. This set of dataestablishes some important parameters ofthe interlingua between pictures and theirnames.

Second, we had subjects rate the degreeof agreement between their mental imageof a concept and its name by presentingsubjects with the most common name ofeach picture, asking them to image itspictorial form, and then presenting themwith the picture and asking them to ratethe degree to which the picture resembledtheir image. This set of data establishessome parameters of the interlingua betweennames and their images.

Because familiarity has been shown tohave such important effects on variousmemory and cognitive processing tasks,we had subjects rate the familiarity ofeach pictured object. Use of the picturerather than its name in this rating isimportant because many concrete nounshave more than one picturable sense (e.g.,nut, top, and pipe). Word frequencycounts are particularly useless for thispurpose because they do not distinguishbetween parts of speech or metaphoricversus concrete usages. For example, welland saw both have very high KuCera-Francis frequencies, primarily because ofnonnoun usages, and both house and doorhave high frequencies, presumably becauseof metaphoric rather than concrete usages.

Finally, because visual complexity mayaffect such variables as naming latencies,tachistoscopic recognition thresholds, andperhaps memorability, we had subjectsrate the visual complexity of each picture.

Method

Subjects

A total of 219 subjects participated. Forty-twosubjects were run in the name agreement task,and different groups of 40 subjects served in theother three major tasks. An additional two groupsof 40 and 17 subjects served in two tasks thatwere supplementary to the main image agreementtask, picture-name agreement and image vari-ability. All were volunteers from introductorypsychology courses who participated to fulfill acourse requirement, except the 17 who participatedin the image variability task who were studentsin an experimental psychology laboratory course.All were native English speakers, and approxi-mately equal numbers of males and females servedin each of the tasks. Subjects were run in smallgroups of from 5 to IS in a classroom.

Stimulus Materials

The pictures (presented in Appendix A) wereblack outline drawings on a white background.The drawings were photographed and made intotransparent slides. Each picture occupied approxi-mately the same subjective amount of space onthe slide (as described previously).

General Procedure

The pictures were projected sequentially on alarge white screen at the front of a slightly darkenedroom using a Kodak Carousel slide projector.Four different random sequences of the 260 slideswere used by constructing two different randompermutations and running each in the forwardand backward order. Approximately equal numbersof subjects in each of the tasks saw each of thefour sequences.

At the start of each task, subjects were read astatement that described the importance of nor-mative data for pictures and encouraged them torespond carefully and consistently. Each slide waspresented for a period of from 3 to 5 sec. Subjectsrecorded their responses on individual data sheets.They were instructed to respond to every slide,leaving no blanks. Halfway through the slides, thesubjects were given a S-min rest period.

Name Agreement

Subjects were instructed to identify each pictureas briefly and unambiguously as possible by writingonly one name, the first name that came to mind.They were told that a name could consist of morethan one word. Specific instructions, designed toelucidate the sources of naming failures, were alsogiven. Subjects were instructed to respond DKO(don't know object) if the picture was of an object


unknown to them. If the object was known butthe name was unknown, they were to respondDKN (don't know name). And, finally, if theyknew the name but it was momentarily irretriev-able, they were to respond TOT (tip-of-the-tongue).

Familiarity

Subjects were asked to judge the familiarity ofeach picture "according to how usual or unusualthe object is in your realm of experience." Fa-miliarity was defined as "the degree to which youcome in contact with or think about the concept."They were told to rate the concept itself, ratherthan the way it was drawn. If they did not knowwhat the object was, they were to respond withthe letters DKO.

A S-point rating scale was used in which 1 indi-cated very unfamiliar and 5 indicated very fa-miliar. In this and all rating tasks, subjects weretold to assign only one whole-number value to eachpicture and were encouraged to employ the fullrange of scale values throughout the set of pictures.To give subjects an idea of the range of familiarityin the set, so as to provide anchors for the scale,the first 30 pictures in the sequence were presentedto them prior to their ratings.

Visual Complexity

Subjects were instructed to rate the complexityof each picture on a 5-point scale in which 1 indi-cated very simple and 5 indicated very complex.Complexity was defined as the amount of detailor intricacy of line in the picture. They were toldto rate the complexity of the drawing itself ratherthan the complexity of the real-life object it repre-sented. As in the familiarity ratings, subjects wereshown the first 30 slides in the sequence to permitthem to anchor their scales.

Image Agreement

Subjects were asked to judge how closely eachpicture resembled their mental image of the object.At the start of the session, subjects were informedof the general nature of the pictures, that is, thatthey were relatively simple black-and-white outlinedrawings. Prior to presenting each picture, theexperimenter called out the picture's most commonname (as determined from data of the name agree-ment task), waited approximately 3 sec, and thenprojected the picture on the screen. During the3-sec period, subjects looked at the darkened screen(or closed their eyes) and formed a mental imageof the object named. Following the appearance ofthe picture on the screen, subjects rated the degreeof agreement between their image and the pictureusing the S-point scale. A rating of 1 indicatedlow agreement, that the picture provided a poormatch to their image, and a rating of 5 indicatedhigh agreement.

Subjects were instructed to write the letters NI(no image) if they could not form an image of an

object for any reason. The NI response did notdistinguish between two possible sources of imagefailure: not knowing the object to which the namereferred or knowing the object but being unableto image it. If subjects imaged a different objectfrom the one pictured (e.g., imaging a lump ofmetal to the name "iron," instead of a householdappliance), they were to respond DO (differentobject).

Supplementary Image Tasks

Picture-name agreement. In this variation of theimage agreement task, the experimenter called outthe name of the concept concurrently with pre-senting its picture. Subjects were asked to ratehow closely the picture matched the way theyexpected the object to look. Thus, in this variationsubjects were not explicitly instructed to form amental image and hence were evaluating the matchbetween a picture and its name. The same S-pointscale was used in this task as in image agreement.

Image variability. In this variation of the imageagreement task, subjects were given only the namesof the concepts and asked to evaluate the potentialvariability in images to a particular name byrating how many different images each name couldevoke. The ratings were made on a S-point scalein which 1 indicated that the name evoked fewimages and S indicated that the name evoked manydifferent, and different appearing, images.

Results and Discussion3

The pictures are presented in Appendix Ain alphabetical order according to themost frequent names assigned to them bythe subjects. Each picture is also numberedfor convenient reference to the table ofnorms presented in Appendix B. AppendixC presents more detailed information onthe nature of the difficulties that subjectsencountered when naming, imaging, orrating the familiarity of the concepts.

Appendix B presents the following in-formation for each picture: the identifyingnumber and most frequent name; twomeasures of name agreement, the informa-tion statistic H and the percentage ofsubjects giving the most common name;and the means and standard deviations forratings of image agreement, familiarity,and visual complexity. Name and imagefailures were not included when computingthe ratings. The Kucera-Francis (1967)

3 All effects reported as significant are at the .05level or better.


frequency counts are also included inAppendix B for the 240 concepts withsingle-word names. A frequency of zeroindicates that the single-word name didnot occur in the KuCera-Francis corpus.Because the Ku6era-Francis norms do notprovide counts for names of more than oneword, these unavailable frequencies are in-dicated by a dash. Counts for alternatespellings of a name were combined onlywhen both spellings occurred in our subjects'naming responses (e.g., ax and axe werecombined, but cigarette and cigaret werenot). Either the singular or plural KuSera-Francis count was used, the one correspond-ing to the sense of the picture (e.g., glass forwater glass and glasses for eyeglasses).

The Carroll- White (1973a) age-of -acqui-sition norms are given in Appendix B forthe 89 names for which they were available.The age-of-acquisition norms were obtainedby having adult subjects estimate the ageat which they first learned the words andthen converting these average estimates toa 9-point scale in which 1 represents an ageof 2 years and 9 represents an age of 13years and above.

The information statistic H was com-puted for each picture by the formula

H =

where k refers to the number of differentnames given each picture and pi is theproportion of subjects giving each name.The three categories of naming failures —DKOs (don't know object), DKNs (don'tknow name), and TOTs (tip-of-the-tongue)— were eliminated when computing Hvalues, but not when computing the per-centage agreement scores. Thus, a picturewith H value of .0 can have a percentageagreement score that is less than 100% be-cause the picture produced naming failuresin some subjects.

Because so many of the concepts showedperfect name agreement (i.e., an H valueof .0), we used a strict criterion for countingdifferent instances of names. In many cases,the name given by a subject was similar tobut not identical to an established name

category. These cases included misspellings,abbreviations, elaborations, and multiplenames. Misspellings were included with thecorrectly spelled name when computing H,commonly accepted abbreviations such asTV for television were counted as separatenames, uncommon abbreviations such asb. carriage for baby carriage were includedwith the unabbreviated form, elaborationssuch as bell pepper for pepper were countedseparately, and when a subject wrote downtwo distinct names for a picture, only thefirst was counted.

A picture that elicited the same namefrom every subject in the sample who wasable to name it has an // value of .0 andindicates perfect name agreement. An itemthat elicited exactly two different nameswith equal frequency would have an Hvalue of 1.00. Increasing H values indicatedecreasing name agreement and, generally,decreasing percentages of subjects who allgave the same name (the correlation be-tween H and percentage agreement acrossthe set is — .94). However, the H value cap-tures more information about the distribu-tion of names across subjects than does thepercentage agreement measure. For ex-ample, if two concepts both are giventheir dominant name by 60% of thesubjects, but one is given a single othername and the second is given four othernames, both concepts will have equalpercentage agreement scores, but the firstwill have a lower H value. Accordingly, weshall use the PI value as the primarymeasure of name agreement in subsequentanalyses.

Table 2 presents summary statistics forsix of the measures shown in Appendix B.The 25th (Ql) and 75th (Q3) percentilesare shown to facilitate selection of conceptsfrom the extremes of the distribution.

The distribution of H values has a lowmean and is positively skewed, reflectingthe fact that many concepts show highname agreement (54 concepts have Hvalues of .0, and 85 have H values of .16or below, where .16 represents consensusamong all but one of the subjects on a


Table 2Summary Statistics for all the Variables

H IA K-F K-F'» A-A

MSD

MdnQiQ3IQRRange

(limits)Skew"

.558

.526

.42

.12

.87

.750-2.551.50

3.69.585

3.723.274.15.88

2.05-4.73

.96

3.29.956

3.322.494.091.601.18-4.90

.93

2.96.897

2.932.283.591.311.00-4.781.02

37.8688.09

133

33300-

8972.00

2.451.59

2.351.443.211.770-9.64

.95

3.091.03

2.932.293.891.601.34-5.481.50

Note. H = name agreement; IA = image agreement; F = familiarity; C = visual complexity; K-F= frequency; K-F' = transformed frequency; A-A = age of acquisition; Ql = 25th percentile;Q3 = 75th percentile; IQR = interquartile range."Skew = (Q3 - Mdn)/(Mdn - (21); >1 is positively skewed.bK-F' = (K-F)».

picture's name). Of the three measuresbased on ratings, both familiarity andcomplexity show a greater range of valuesthan image agreement, reflecting greaterconsensus among subjects on the extremesof the scale. Complexity ratings are fairlysymmetric around the midpoint scale valueof 3, whereas both familiarity and imageagreement ratings tend to be negativelyskewed, reflecting the fact that few con-cepts were judged to be very low ineither familiarity or image agreement. Forimage agreement, it is rare for subjectsto agree that their visual image does notmatch the picture, since the lowest IArating was 2.05.

The Kuc"era-Francis frequencies are posi-tively skewed, because of the few conceptswith very high frequencies. Because wewere particularly interested in the compar-ison between Ku&ra-Francis and familiar-ity, we sought a simple transformation ofthe Kucera-Francis values that wouldhave the properties that zero frequencies bedenned (which is not true of the logtransform), and that the distribution oftransformed scores show a negative skewcomparable to that obtained for familiarityratings. The cube root transformation(K-F') met both requirements and isused in all subsequent analyses.

Correlations Among the Measures

The characteristics of pictures we choseto measure were selected for two reasons.First, they appeared to be importantvariables in the various semantic andepisodic memory tasks in which they havebeen used, and second, they appeared torepresent distinctly different attributes ofthe pictures. To determine the degree ofrelationship among our measures, andbetween them and some important mea-sures already available, we computed theinteritem correlations among all attributespresented in Appendix B (with the excep-tion of percentage agreement). Kucera-Francis frequencies were available for only240 of the 260 concept names—those forwhich the name was a single word. Becauseage-of-acquisition measures were availablefor only 89 of the concepts, a separate setof correlations was computed for thissubset. Table 3 presents the matrix ofsignificant correlations for all concepts (orfor the subset of 240 for the correlationsinvolving K-F'), and Table 4 presents thematrix for the subset of items havingage-of-acquisition values. With a singleexception, the correlations among the fourmeasures collected in the present study areall relatively small in magnitude, suggestingthat the measures do indeed represent


Table 3Significant Correlations Among the Measured Variables for all 260 Concepts andBetween Those Variables and Transformed Kucera- Francis Frequency for 240 Concepts

Variable 1 2 3 4

1. Name agreement2. Image agreement3. Familiarity4. Visual complexity5. Transformed frequency

1.000-.313 1.000

— .138 1.000.130 -.209 -.466— — .363

1.000-.180 1.000

independent attributes. The single excep-tion is the high negative correlation betweenfamiliarity and visual complexity, which weshall discuss further in some detail.

We expected that, of the measuresobtained in the present study, familiaritywould show the highest correlation withboth frequency and age of acquisition. Asshown in Tables 3 and 4, familiarity ispositively correlated with frequency (.363and .499 for the set of 240 and 89 concepts,respectively), although the correlationsare modest in size and even smaller whenthe untransformed frequency values areused (.156 and .372 for the set of 240 and89 concepts, respectively). Presumably,the reason that the correlation is so lowis because frequency values do not distin-guish between different meanings of a word.The fact that the correlation is higher forthe subset of concepts with age-of-acquisi-tion measures suggests that the wordsselected for inclusion in the age-of-acquist-tion norms were low in ambiguity ofreference.

Familiarity is negatively correlated withage of acquisition (— .550), meaning thatnames of concepts that are familiar also

are learned at an early age. The correlationbetween age of acquisition and transformedfrequency, although significant, is some-what lower (— .482) than that found byCarroll and White (1973a), probably be-cause their range on both variables wasgreater than in the present study.

Of the correlations among our ownmeasures, name agreement and imageagreement are negatively correlated ( — .313and —.366), suggesting that concepts thathave many names also evoke many differentimages. Image agreement is positively(although modestly) correlated with famil-iarity (.138), although the correlation isnot significant for the subset of 89 concepts.Objects that are familiar tend to produceimages that agree with the pictured form.

Visually complex pictures tend to havemany names, tend to show low imageagreement (although only for the entireset), and tend to be rated as unfamiliar.Since the visual complexity of a picturedepends on how it is drawn, one interpreta-tion of the effects of visual complexity onother attributes is that complexity isentirely a product of the whim of theartist's pen. Under this interpretation,

Table 4Significant Correlations A mong all Variables for 89 Concepts forWhich Age-of-Acquisition Norms Were Available

Variable 1

1. Name agreement2. Image agreement3. Familiarity4. Visual complexity5. Transformed frequency

1.000-.366 1.000

— —.230 —— —

1.000-.413

.4991.000

-.214 1.0006. Age of acquisition -.550 .233 -.482 1.000


complexity correlates positively with Hbecause complex drawings are more difficultto recognize, negatively with image agree-ment because mental images are simplerthan complex drawings, and negativelywith familiarity because complex drawingsare more novel than simple drawings.

An alternative interpretation of theeffects of visual complexity is that theamount of detail in a picture is determinedprimarily by certain characteristics of theobject itself rather than by the artist's style.We explicitly had the concepts drawn sothat the amount of pictorial detail wouldbe consistent either with the amount ofdetail in the real-life object or the amountof detail in its conventional representation.According to this interpretation, the twosources of the effects of visual complexityon other variables are, first, propertiesinherent in the object itself—how complexit is in real life—and second, propertiesinherent in its grammar of representation.The correlation between complexity andimage agreement most likely stems fromthe first source, the complexity of thereal-life object. That is, a complex objectmay be imaged in a greater variety of waysthan a simple object because the formerhas more possible visual features than thelatter. Thus, a complex object's image isless likely to match any particular pictorialrepresentation. In contrast, the correla-tions of complexity with familiarity andname agreement most likely stem from thesecond source, the conventions developedfor the concept's representation. Familiarobjects, which by definition must bespoken and thought about frequently,have necessarily developed efficient andsimplified verbal and visual codes. Thus,they are named briefly and consistentlyand pictured simply and consistently.Their simplified grammar of reference andof representation thereby accounts for thecorrelation between complexity and nameagreement and between complexity andfamiliarity.

Support for the simplified visual codeviewpoint is provided by data from anearlier set of 165 picture stimuli drawn bya different artist (one of the authors), for

which a similar set of norms was collected(Vanderwart & Snodgrass, Note 2). Al-though the 165 concepts from the previousstudy were a subset of the present set,their style of drawing differs considerablyfrom the present set. Yet, the pattern ofcorrelations between visual complexity andboth familiarity and name agreement isvirtually identical. Complexity was nega-tively correlated with familiarity, as inthis study ( — .363 vs. —.466 here), andpositively correlated with name agreement(.199 vs. .130 here).

Naming and Imaging Failures

Although the major purpose of thepresent study was not to elicit namingand imaging failures from our subjects(indeed, quite the opposite was desired),it is of interest to determine whether thereis some consistency from task to taskabout which concepts elicited failures, andwhether any of the measured character-istics of the pictures are related to suchfailures. Appendix C presents all conceptsfor which any naming or imaging diffi-culty occurred.

As indicated earlier, subjects in thenaming task indicated, when they wereunable to name a picture, whether theydid not know the object, did not knowthe name, or were in a tip-of-the-tonguestate. Subjects in the image agreementtask indicated when no image had oc-curred, and subjects in the familiarityrating study indicated when they did notknow the object pictured. The highestrate of failure occurred in the nameagreement task (1.7% across the threecategories), the next highest in the imageagreement task (.7%), and the smallestin the familiarity task (.3%). It is, ofcourse, not surprising that the highestrate of failures occurred in naming, sincesubjects in that task were required toproduce a name or indicate the reasonwhy they could not. In the rating tasks,on the other hand, subjects could moreeasily ignore the instructions and rate thepicture whether or not they recognized


the object pictured or failed to generatean image to its name.

Because naming and imaging failuresoccurred so infrequently, it seemed desir-able to combine responses across taskswhenever possible. Accordingly, phi coeffi-cients were computed across similar re-sponse categories to determine whetherthere was significant agreement acrosstasks or responses about which particularconcepts led to naming or imaging failures.There was significant agreement betweenthe naming and familiarity groups as towhich concepts received one or more DKOresponses ($ = .439), so the DKO re-sponses from these two groups were com-bined. There was also significant agreementbetween concepts for which one or moreDKN and TOT responses were given inthe naming task ($ = .400), so these re-sponses were combined and all classifiedas TOTs. This seemed a reasonable clas-sification, since subjects may have haddifficulty discriminating between not know-ing an object's name and not being ableto retrieve it. Combining responses in thisway produced three categories of responses:DKOs (from naming and familiaritygroups); TOTs (from TOTs and DKNsin the naming groups); and NIs from theimage agreement group. The intercorrela-tions among these three groups of responseswere also significant, but lower. (For DKOversus TOT, $ = .242; for DKO versusNI, $ = .163; and for TOT versus NI,$ = .290.)

To determine on which, if any, variablesthe difficult-to-name or difficult-to-imageconcepts differed from the rest, those con-cepts for which either three or more sub-jects gave responses falling into one of thethree combined categories, DKO, TOT,and NI, or whose sum on all such responsesexceeded four, were identified as difficultto name/image. Twenty-nine concepts fellinto this category. The 29 concepts showinghigh naming/imaging failures from highestincidence to lowest (concepts within pa-rentheses are tied) are artichoke, chisel,spinning wheel, asparagus, French horn,cloud, (orange, raccoon, watering can,wrench), (nut, ostrich), (ant, peach, pea-

cock, pepper, rhinoceros, rolling pin, sheep),(accordion, beetle, celery, cherry, harp,potato), and (ironing board, plug, seahorse, thimble).

The mean values for this subset of 29on each of the four characteristics mea-sured in the present study, and on trans-formed frequency values, were comparedwith the mean values for the remainingconcepts. The subset showed significantlyhigher H values, <(258) = 3.92, est <rM

= .101; significantly lower familiarityratings, *(258) = 4.82, est <TM = .178; andsignificantly lower K-F' values, i(238)= 3.03, est <TM = .344, whereas neitherimage agreement nor complexity ratingswere significantly different between thetwo subsets.

The fact that items with higher H valuesand lower familiarity or frequency valuesare more difficult to recognize, name, orimage is consistent with previous researchon picture-naming latencies. However, ourstimuli were not purposely designed toelicit naming or imaging failures (to thecontrary, in fact), so this post hoc analysisshould be regarded cautiously.

Sources of Name Disagreement

To determine what kinds of names thatsubjects give to concepts when they dis-agree with the concept's dominant name,we classified nonmodal names into thefollowing classes: synonyms, coordinates,superordinates, subordinates, and other.Synonyms included a modifier added tothe basic name that was redundant withthe pictured concept (e.g., green pepperfor pepper and bunch of grapes for grapes).Coordinates were defined as different ex-emplars of the same category (e.g., spiderfor ant, avocado for artichoke, and an-telope for deer). Superordinates includedinsect or bug for ant, fruit for peach, andbird for chicken. Subordinates were definedas a subclass of the concept pictured, andincluded polar bear for bear, sparrow forbird, and rose for flower.

Table 5 presents the results of thisclassification for 13 selected categories, ofwhich 12 are the Battig and Montague


Table 5Percentage of Name Responses in 13 Selected Categories for (a) Dominant Responses;(b) Synonyms; (c) Coordinates; (d) Superordinates; (e) Subordinates;and (/) Naming Failures

Category

Insects (8)Musical instruments (9)Vegetables (13)Birds (8)Carpenter's tools (12)Clothing (19)Fruits (11)Animals (30)Kitchen utensils (14)Basic level (4)Vehicles (10)Human body parts (12)Furniture (14)

(a)Dom

76858384878991908595858884

(b)Syn

0044231390

131015

(a + b)

76858788899292939495989899

(c)Coord

11658653510001

(d)Super

9212011030001

(e)Sub

0200020015220

(f)Fail

4463504210000

Number of exemplars in each category are in parentheses.

categories. All exemplars listed in Table 1are included except those for which thepictured form of the concept is judged todiffer from the meaning intended (markedwith a dagger in Table 1) or in which anexemplar appears in another category withhigher frequency (marked with an asteriskin Table 1). Three of the Battig andMontague categories are omitted eitherbecause of the small number of exemplarsin each or because so many of the picturedconcepts were not the meaning intended.These omitted categories are weapons,building parts, and toys. An additionalcategory has been added to Table 5, thebasic level category consisting of the con-cepts bird, fish, flower, and tree.

Of particular interest in Table 5 is thecolumn labeled (a + b), which representsthe total percentage of dominant andsynonymous names in each category andthus can be considered to be the per-centage of correct names for the concept.Percentage of correct names varies froma low of 76% for the insect category toa high of 99% for the furniture category.

The categories differ by the kinds ofincorrect names given, with insects elicitinga high percentage of coordinates andsuperordinates, vegetables eliciting a highpercentage of naming failures, and the

basic level category the only one to showany appreciable number of subordinateresponses (although these occurred ex-clusively for the concepts of bird andflower).

Concept Versus Name Agreement

As we indicated previously, the criterionfor denning name agreement was verystrict. In particular, such synonymousterms as TV for television and purse forpocketbook were treated as separate namecategories in computing both H and per-centage agreement. Although this defini-tion of name agreement is useful as ameasure for predicting such cognitive tasksas naming latencies, it is less useful forsuch tasks as picture recall, in whichsynonyms for the dominant name wouldprobably be scored as correct responses.

Accordingly, we have identified conceptswhose high values of H reflect linguisticambiguity, as opposed to conceptual orpictorial ambiguity. As we use the term,linguistically ambiguous items are thosethat elicit a variety of names that areapproximately synonymous and thus maybe said to have high concept, as opposedto name, agreement.


Table 6Items of High Concept Agreement and Low Name Agreement

Identifyingno.

213171929334755717479

8898

101105112127134136146

152153161178184189191214219221228235239243258

Name

AirplaneBaby carriageBarnBaseball batBlouseBowCarChickenDeerDollDresser

FingerFoxFrying panGlassesGunKettleLegLeopardMoon

Nail fileNecklacePaintbrushPocketbookRecord playerRoller skateRoosterSpool of threadStoveSuitcaseTelevisionToeTraffic lightTrumpetWine glass

% Conceptagreement

889890

10090959090839390

88889598

10098888695

9893959090

10088989890

100959090

100

Synonyms(in order of frequency)

Plane, jetCarriageBarn and SiloBatShirt, jacketRibbon, bow tieLincolnHenAntelopeBaby, little girlBureau, chest,

chest of drawers, drawersIndex fingerWolfPanEye glassesPistol, revolverTea kettle, teapotKneeTigerQuarter moon, crescent

moon, half moonFilePearl necklace, beads, pearlsBrushHandbag, pursePhonograph, turntableSkateChickenThread, spoolOven, rangeLuggageTV, television setBig toe, toesStop lightHornGlass, goblet

High concept agreement pictures weredenned as those items whose H valueswere greater than or equal to 1.00 andfor which three or fewer subjects gavenames that were not synonyms of thedominant name. Of the 48 concepts havingH values of 1.00 or greater, 35 fulfilledthe criterion of high concept agreement.These concepts are listed in Table 6, alongwith the recomputed percentage agreementscore when all synonyms are consideredequivalent to the dominant name. Thetable also lists the synonyms in order of

the frequency with which three or moresubjects gave them as responses. Nameswere classified as synonyms on the basisof the experimenters' judgment and withrespect to the picture's appearance. Thus,the name baby was considered reasonablysynonymous with the pictured doll, al-though it would not have been had arag-type doll been pictured. In any case,the information provided in Appendix Cis sufficiently detailed to permit an in-vestigator to recalculate the data shownin Table 6. The average percentage con-


cept agreement score is 93% for these35 concepts compared to the average of64% name agreement.

Supplementary Image Agreement Measures

There were two subsidiary image agree-ment measures obtained for the s'et ofpictures: picture-naming agreement, ob-tained by having subjects judge the degreeto which a presented picture matched itsname, and image variability, obtained byhaving subjects rate how many differentimages a name could have in the absenceof the picture. These supplementary mea-sures were obtained to clarify the data onimage agreement, which seems to be themost subjective judgment of all the mea-sures. We expected the correlation betweenimage agreement and picture-name agree-ment to be high and positive, and thecorrelation between image agreement andimage variability to be lower and negative.

Picture-name agreement. As expected,the correlation between image agreementand picture-name agreement was highlysignificant across the 260 picture-namepairs (r = .739), suggesting that the processsubjects used to decide whether an imagegenerated to a name agreed with a picturewas not much different from that used todecide whether a name agreed with apicture. However, the mean ratings weresignificantly higher for picture-name agree-ment than for image agreement, <(259)= 9.20, est OM = .036, either because thepicture immediately clarifies the meaningof the name or because the presence ofthe picture biases the subject toward theparticular appearance of the object shown.Obviously, subjects must access some in-formation in long-term memory about howobjects look in order to make picture-nameagreement judgments, even though theymay not actually form a mental image todecide that a picture matches its name.

Image variability. The correlation be-tween image agreement and image vari-ability was negative and lower, as ex-pected (r = —.441). Our expectation thatthe two measures should be negativelycorrelated was based on the notion that

a mismatch between a particular subject'simage of an object and its picture in theimage agreement task is diagnostic of thepossibility that a particular concept canhave many different images. Conversely,if a subject can create many different anddifferent-appearing images for a concept,that concept's picture is less likely tomatch any one of his or her images. Forexample, such concepts as artichoke, lemon,and zebra, which were rated lowest onimage variability, are examples of con-cepts that look very similar from instanceto instance, whereas car, dog, dress, andhouse, which were rated highest on imagevariability, are examples of concepts thatlook very different from instance to in-stance. Interestingly, all four of the basiclevel concepts, bird, fish, flower, and tree,had image variability ratings in the upper4% of the distribution.

Significance of the Normative Data forFuture Research

The intercorrelations among name agree-ment, familiarity, visual complexity, andimage agreement are quite low, suggestingthat the four measures represent largelyindependent attributes of the pictures. Itwill be useful to consider the possibleeffects of these attributes with regard tofive experimental paradigms widely usedin studies of picture-word processing. Thefirst two are semantic memory processingtasks, and the remaining three are episodicmemory tasks. The semantic memory tasksare picture naming, in which the latencyto name a pictured object is measured,and picture categorization, in which thelatency to classify an object into itssemantic category is measured. The epi-sodic memory tasks include two types ofrecognition memory tasks as well as theclassical free-recall task. The two recogni-tion memory tasks are stimulus recogni-tion, in which the form of the item (pictureor word) is the same in the test as it isin the study presentation, and conceptrecognition, in which the study and testforms of the item are different, for example,a picture study item is tested as a wordand vice versa.


Table 7Means of the Battig and Montague Ranks, Familiarity Ratings, and Complexity Ratingsof Relevant Exemplars of 12 of the Battig and Montague Categories, With the CorrelationBetween Rank and Familiarity Within Each Category

Category (n)» Rank Familiarity Complexity

Fruits (11)Insects (8)Musical instruments (9)Human body parts (12)Carpenter's tools (12)Vehicles (10)Clothing (19)Vegetables (13)Kitchen utensils (14)Furniture (14)Animals (30)Birds (8)

7.67.78.4

12.011.312.515.818.920.523.824.029.7

3.362.442.604.663.233.583.923.184.164.102.362.16

-.642*-.502-.611*-.198-.120-.904**-.395*-.623*-.513*-.515*-.305*-.473

2.314.044.042.422.373.812.613.032.262.663.843.82

M 16.0 3.31 -.483 3.10

• Number of exemplars per category in parentheses.*p < .05. **p < .01.

Name Agreement

The variable of name agreement is likelyto affect naming latencies for pictures.The results of Lachman (1973) would leadus to expect that concepts with high Hvalues will have longer naming latenciesthan concepts with low H values. How-ever, we would not expect name agreementto affect categorization times for picturesif, as is commonly believed, subjects cate-gorize a picture without first accessing thename. Thus, for example, pictures fromthe insect category, many of which aredifficult to name, should be as quicklycategorized as pictures from the furnitureor clothing category, most of which areeasy to name.

Similarly, name agreement is unlikelyto influence stimulus recognition memoryto any great extent, since picture recogni-tion appears to be mediated largely byvisual rather than by verbal codes (Nelsonet al., 1976; Paivio, 1971; Snodgrass,Burns, & Pirone, 1978). On the otherhand, in a concept recognition task, nameagreement should be a relevant variable,since a picture with many possible nameswill have a verbal code that is unlikely tomatch the experimenter-determined name.Accordingly, concepts with high name

agreement will be better recognized thanconcepts with low name agreement in aconcept recognition memory paradigm.

In a recall task, it is obvious that nameagreement will affect recall of pictures,since recall requires retrieving the nameof the concept. Pictures that have highH values either are named with difficultyor have many synonymous names. Picturesthat are named with difficulty will probablybe recalled less well, although pictureswith many synonymous names may berecalled as well as pictures with uniquenames. Table 6 provides information neces-sary to distinguish these two classes ofhigh H value pictures and also lists thesynonyms that are likely to appear as"correct" recall responses.

Familiarity

The familiarity rating of a picture isanalogous to the frequency count of theword form of the concept, and the twoare highly correlated. It is well-knownthat an inverse relation exists betweennaming latency and the frequency of thepicture's name. Thus we would expectpictures with low familiarity to havelonger naming latencies than pictures withhigh familiarity. The form of the famili-


arity function may be somewhat differentfrom the frequency function because thefrequency count includes all the meaningsof the word form.

Category judgments are made morequickly for concepts that are prototypicalof their category, in the Roschian sense.Accordingly, if familiarity and prototypi-cality are correlated, then it follows thatfamiliar concepts will also be categorizedmore quickly. It seems likely that whentypicality is denned as the frequency withwhich the name of an exemplar is retrievedfrom semantic memory (as we have donefor the Battig and Montague categories),more familiar concepts should be retrievedmore readily, and hence have a higherrank than unfamiliar concepts.

To investigate the relation between typi-cality and familiarity, we computed cor-relations between rank and familiarity for12 of the Battig and Montague categories(using the same criteria for exemplar in-clusion employed for Table 5). Table 7presents the correlations for each of the12 categories, with the mean ranks andfamiliarity ratings by category. (The meancomplexity ratings by category are alsoshown and will be discussed.)

Eight of the 12 correlations are signifi-cant and negative, indicating that highlyranked items (those with low numericalranks) also tend to be more familiar. Themean of the 12 correlations ( — .483) isalso significantly different from zero bya / test, / ( l l) = 7.91, est <rM = .061. Thus,the hypothesis that typical category ex-emplars are also more familiar is con-firmed, and hence we expect that familiarpictures will be categorized faster thanunfamiliar pictures.

Table 7 also reveals that the categoryexemplars vary in familiarity, from a lowof 2.16 for the bird category to a highof 4.66 for the category of human bodyparts. Some of this variation is a resultof selection of the exemplars (e.g., all ofour birds are quite atypical). A consider-able amount of variation, however, is dueto the nature of the category itself. Ananalysis of variance performed on the 12mean familiarity values was highly signifi-

cant, F(ll, 148) = 25.71, MSe = .343.A Newman-Keuls test revealed five groupsof categories that were significantly dif-ferent from one another. They are, inorder from lowest familiarity to highestfamiliarity, (a) birds, animals, insects,musical instruments; (b) vegetables, car-penter's tools; (c) fruits, vehicles, clothing;(d) furniture, kitchen utensils; and (e)human body parts.

We would expect picture familiarity toaffect episodic memory in much the sameway as word frequency does. Specifically,the recognition-recall paradox for words—that frequent words are better recalledthan recognized and infrequent wordsbetter recognized than recalled—shouldalso hold for familiar and unfamiliar pic-tures. As we have noted, familiarity is a"purer" measure of the picturable senseof a word than frequency. Thus, famili-arity should be a better predictor ofmemory performance for pictures and forwords in which the experimental contextbiases a particular word meaning. Forexample, the concept recognition task, inwhich words are tested with their pictureform, presumably biases word interpreta-tion toward its picturable noun sense, anda categorized list recall task presumablybiases word interpretation toward the senserelevant to the category (e.g., nut as a toolvs. as a vegetable).

Visual Complexity

The complexity j>( a picture primarilyreflects the superficial visual characteristicsof the object and its conventions of pic-torial representation. The picture-namingtask presumably requires at least twosteps: picture recognition and name re-trieval. The first step of picture recogni-tion may take longer for more complexpictures, so that visual complexity mayaffect naming latencies.

Similarly, categorization also requirespicture recognition as a first step, althoughthe process of name finding may be by-passed. Pictures that vary in complexitywithin a category may show correspondingvariations in categorization latencies. Fur-


thermore, the extent to which categoriesas a whole differ in complexity shouldaffect categorization times between cate-gories. Table 7 shows considerable varia-tion in mean complexity from category tocategory, with the categories of kitchenutensils and fruits having relatively simpleexemplars, and the categories of insectsand musical instruments having relativelycomplex exemplars. An analysis of variancerevealed that the 12 categories did indeedvary significantly in complexity, F(ll, 148)= 20.63, MSe = .339. A Neuman-Keulstest revealed that three groups of cate-gories were significantly different from oneanother. They are, in order from lowestcomplexity to highest complexity, (a)kitchen utensils, fruits, carpenter's tools,human body parts, clothing, and fur-niture; (b) vegetables; and (c) vehicles,birds, animals, musical instruments, andinsects.

In episodic memory tasks, visual com-plexity may affect stimulus recognitionmemory. The intricate detail in a complexpicture (e.g., motorcycle) or utter lack ofit in a simple picture (e.g., star) mayfunction to make the stimulus novel andthus more recognizable than medium com-plex pictures. Some effect of complexitymay also obtain in a concept recognitiontask. One might speculate that mentalimage codes generated to the word formof a concept are relatively simple andschematic. Thus, a very complex objectpresented first in word form and testedin picture form would be less well recog-nized than a simple object because itsschematic image would not provide asgood a match to its detailed picture. Ingeneral, complexity should have its greatesteffects in recognition rather than recall.

Nelson et al. (1976) have shown thatvisual or schematic similarity among a setof pictures reduces the picture superiorityeffect in paired-associate recall, when pic-tures are the stimulus members. Visual com-plexity is different from similarity in thatthe former refers to an aspect of individualpictures, whereas the latter refers to acharacteristic shared by two or more pic-tures. However, the extent to which objects

can be visually similar presumably dependson their overall level of complexity, withsimpler objects (e.g., ball, balloon, apple)having more opportunity for visual simi-larity than complex objects. Thus, com-plexity probably also affects the degree towhich the picture superiority effect can bereduced by introducing visual similarity.

Image Agreement

Image agreement will be likely to in-fluence semantic tasks in some interestingways. As we just noted, the picture-naming process presumably entails twosubprocesses: First, the image code cor-responding to the picture must be accessed(i.e., the object must be recognized forwhat it is), and then the verbal labelmust be accessed. The time consumed bythe second step, name finding, may beheld constant by selecting items of thesame degree of name agreement and fa-miliarity. The first step, image finding,however, is presumably directly dependenton image agreement. A picture with lowimage agreement would take longer toname because its corresponding imagewould take longer to identify. To date,we have unfortunately not been able toanalyze naming latencies in such a fashiondue to the unavailability of image agree-ment norms.

In categorization tasks, image agreementmay act in a way analogous to proto-typicality. The prototypicality of a con-cept as an exemplar of its class is a primedeterminer of superordinate category judg-ment latencies. Image agreement is ananalogous measure to prototypicality onthe basic category level. That 'is, imageagreement measures the typicality of theform of the exemplar—it answers thequestion, How good an exemplar is thepicture of the concept it represents? Ac-cordingly, pictures with high image agree-ment should be categorized faster thanpictures with low image agreement, justas typical exemplars are categorized fasterthan atypical exemplars.

As in the case of complexity, imageagreement is more relevant theoretically


to recognition memory than to recall mem-ory. In stimulus recognition, pictures withlow image agreement should be at a disad-vantage because they are less likely toaccess a strong image code or will conflictwith the memory code stored for thatconcept. However, the effects of imageagreement should be more noticeable inthe concept recognition task. Here, for alow image agreement item, if a subjectreceives the word form first, he or she isless likely to generate an image that agreeswith the picture presented in the test.Thus, low image agreement should producepoor recognition when items are given aswords in the study and as pictures inthe test.

In summary, we have presented an ex-tensive set of pictures whose propertieshave been quantified on several variablesof theoretical importance to memory andperception. It is our hope that investi-gators will take advantage of this stan-dardized set in future research, so thattheoretical questions about differences inprocessing between pictures and wordsmay be investigated in the absence ofpotential artifacts attributable to peculiar,and unmeasured, aspects of pictorialstimuli.

Reference Notes

1. Snodgrass, J. G. Toward a model for picture andword processing. Paper presented at the SecondConference on Processing of Visible Language,Niagara-on-the-Lake, Canada, September 1979.

2. Vanderwart, M., & Snodgrass, J. G. A new setof 165 picture stimuli with norms for namingambiguity, visual complexity, familiarity, andprototypicality. Unpublished manuscript, 1977.(Available from J. G. Snodgrass, Department ofPsychology, New York University, 6 WashingtonPlace, New York, New York 10003.)

References

Anderson, J. R., & Bower, G. H. A propositionaltheory of recognition memory. Memory & Cogni-tion, 1974, 2, 406-412.

Battig, W. F., & Montague, W. E. Categorynorms for verbal items in 56 categories: A rep-lication and extension of the Connecticut categorynorms. Journal of Experimental Psychology Mono-graph, 1969, 80 (3, Pt. 4).

Bousfield, W. A., Esterson, J., & Whitmarsh, G. A.The effects of concomitant colored and uncoloredpictorial representations on the learning ofstimulus words. Journal of Applied Psychology,1957, 41, 165-168.

Brown, A. S. Catalog of scaled verbal material.Memory & Cognition, 1976, 4, 1S-45S.

Carroll, J. B., & White, M. N. Age-of-acquisitionnorms for 220 picturable nouns. Journal ofVerbal Learning and Verbal Behavior, 1973, 12,563-576. (a)

Carroll, J. B., & White, M. N. Word frequencyand age of acquisition as determiners of picture-naming latency. Quarterly Journal of Experi-mental Psychology, 1973, 25, 85-95. (b)

Cattell, J. M. The time it takes to see and nameobjects. Mind, 1886, 11, 63-65.

Durso, F. T., & Johnson, M. Facilitation in namingand categorizing repeated pictures and words.Journal of Experimental Psychology: HumanLearning and Memory, 1979, 5, 449-459.

Fraisse, P. Recognition time measured by verbalreaction to figures and words. Perceptual &f MotorSkills, 1960, //, 204.

Intraub, H. The role of implicit naming in pictorialencoding. Journal of Experimental Psychology:Human Learning and Memory, 1979, 5, 78-87.

Kucera, H., & Francis, W. N. Computationalanalysis of present-day American English. Provi-dence, R.I.: Brown University Press, 1967.

Lachman, R. Uncertainty effects on time to accessthe internal lexicon. Journal of ExperimentalPsychology, 1973, 99, 199-208.

Martin, E. Generation-recognition theory and theencoding specificity principle. Psychological Re-view, 1975, 82, 150-153.

Nelson, D. L., Reed, V. S., & McEvoy, C. L.Learning to order pictures and words: A modelof sensory and semantic encoding. Journal ofExperimental Psychology: Human Learning andMemory, 1977, 3, 485-497.

Nelson, D. L., Reed, V. S., & Walling, J. R.Pictorial superiority effect. Journal of Experi-mental Psychology: Human Learning and Memory,1976, 2, 523-528.

Nickerson, R. S. Short-term memory for complexmeaningful visual configurations: A demonstra-tion of capacity. Canadian Journal of Psychology,1965, 19, 155-160.

Oldfield, R. C., & Wingfield, A. Response latenciesin naming objects. Quarterly Journal of Experi-mental Psychology, 1965, 17, 273-281.

Paivio, A. Imagery and verbal processes. NewYork: Holt, Rinehart, & Winston, 1971.

Paivio, A. Mental comparisons involving abstractattributes. Memory & Cognition, 1978, 6, 199-208.

Paivio, A., & Csapo, K. Picture superiority in freerecall: Imagery or dual coding? Cognitive Psy-chology, 1973,' 5, 176-206.

Paivio, A., Rogers, T. B., & Smythe, P. C. Whyare pictures easier to recall than words? Psy-chonomic Science, 1968, 11, 137-138.


Pellegrino, J. W., Rosinski, R. R., Chiesi, H. L.,& Siegel, A. Picture-word differences in decisionlatency: An analysis of single and dual memorymodels. Memory & Cognition, 1977, 5, 383-396.

Potter, M. C., & Faulconer, B. A. Time to under-stand pictures and words. Nature, 1975, 253,437-438.

Potter, M. C., Valian, V. V., & Faulconer, B. A.Representation of a sentence and its pragmaticimplications: Verbal, imagistic, or abstract?Journal of Verbal Learning and Verbal Behavior,1977, 16, 1-12.

Reder, L. M., Anderson, J. R., & Bjork, R. A.A semantic interpretation of encoding specificity.Journal of Experimental Psychology, 1974, 102,648-656.

Rosch, E., Mervis, C. B., Gray, W. D., Johnson,D. M., & Boyes-Braem, P. Basic objects innatural categories. Cognitive Psychology, 1976, 8,382-439.

Shepard, R. N. Recognition memory for words,sentences, and pictures. Journal of Verbal Learningand Verbal Behavior, 1967, 6, 156-163.

Smith, M. C., & Magee, L. E. Tracing the timecourse of picture-word processing. Journal ofExperimental Psychology: General, in press.

Snodgrass, J. G., Burns, P. M.. & Pirone. G. V.

Pictures and words and space and time: Insearch of the elusive interaction. Journal of Ex-perimental Psychology: General, 1978, 107, 206-230.

Snodgrass, J. G., & McClure, P. Storage andretrieval properties of dual codes for picturesand words in recognition memory. Journal ofExperimental Psychology: Human Learning andMemory, 1975, 1, 521-529.

Snodgrass, J. G., Wasser, B., Finkelstein, M., &Goldberg, L. B. On the fate of visual and verbalmemory codes for pictures and words: Evidencefor a dual coding mechanism in recognitionmemory. Journal of Verbal Learning and VerbalBehavior, 1974, 13, 27-37.

Standing, L., Conezio, J., & Haber, R. N. Per-ception and memory for pictures: Single-triallearning of 2,560 visual stimuli. PsychonomicScience, 1970, 19, 73-74.

Tulving, E. Episodic and semantic memory. InE. Tulving & W. Donaldson (Eds.), Organizationof memory. New York: Academic Press, 1972.

Tulving, E., & Thomson, D. M. Retrieval processesin recognition memory: Effects of associativecontext. Journal of Experimental Psychology,1971, 87, 116-124.

Index to Appendixes

Appendix Page

ABC

197205210


Appendix A

The 260 pictures are shown below with their identifying number. They are arranged inalphabetical order according to each picture's most common name. The names of each pictureand their norms are shown in Appendix B with the identifying numbers.

10

12 14

16 18 20

21 23 24 25

26 28

198

31

JOAN GAY SNODGRASS AND MARY VANDERWART

32 R 33 _ 34 35

36

41 42 44 45

49 50

52 53 54 55

56 57 1m

58

^K^sr^^

59 60

t \

61 63 64 65

\

u

NORMS FOR PICTURE STIMULI

67 __—^ 68 70

199

76 80

87 88 90

94

97


102 103 Jg{ 104 105

110

111 112 114

116 118 120

121 122

126 128

Co129 130

133 135


1391

201

141

146 148

151 152 155

156 159

161 163

166 170

202

171


172 173 M\fi^ 174 175

176 177 179 180

181 182 183 184 A 185

187 188 189 190

191 192 194 195

196 197 198 199 200

201 202 203 204 205


207 208

203

210210

211 214

217 218 219 220

221

j223 224 225 n

227 229 230

232 233 234 235

236 237 238 239


242 243 244 245

246 248 249

252

256 258 f 259 260


Appendix B

The following information is shown below: the identifying number of each picture; its mostcommon name; two measures of name agreement, H and percentage agreement; the meanand standard deviation for ratings of image agreement, familiarity and complexity; the KuCera-Francis frequency counts (K-F) for each single-word name; and the Carroll-White age-of-acquisition norms (A-A) for the 89 concepts for which they were available.

Table BlNorms

1.2.3.4.5.6.7.8.9.

10.11.12.

13.14.15.16.17.18.19.20.21.22.23.24.25.26.27.28.29.30.31.32.33.34.35.36.37.38.39.40.41.

42.43.44.

Concept

AccordionAirplaneAlligatorAnchorAntAppleArmArrowArtichokeAshtrayAsparagusAxe

Baby carriageBallBalloonBananaBarnBarrelBaseball batBasketBearBedBeeBeetleBellBeltBicycleBirdBlouseBookBootBottleBowBowlBoxBreadBroomBrushBusButterflyButton

CakeCamelCandle

Name

H

.181.77.54.17.64.16.53.16

1.54.00

1.27.53

1.00.44.00.00

1.31.00

1.00.60.53.00

1.652.18.00.16.53.69

1.89.00.69.28

1.25.17.80.84.00.88.00.00.16

.84

.00

.00

agree

%

886088938198909852

1006990

5293

10010069

100529088

1006050

10098888843

100889574958883

10083

10010098

8395

100

Image agree

M

3.403.233.984.322.924.053.952.273.443.203.504.50

3.652.844.334.423.224.314.582.623.623.652.782.052.924.053.403.332.804.332.282.852.673.792.904.024.353.204.083.924.48

3.453.923.85

SD

1.041.12.85.85

1.24.87.89

1.291.471.051.26.63

1.011.191.18.70

1.111.14.70

1.231.09.99

1.171.13.94

1.011.091.141.091.00.96

1.221.25.89

1.181.06.73

1.271.01.85.92

1.12.99.76

Familiarity

M

2.153.781.651.602.623.984.753.382.293.562.682.28

2.723.202.583.652.382.023.682.181.984.722.681.882.204.123.783.624.184.753.383.722.254.182.884.403.423.804,502.923.85

4.022.083.08

SD

1.20.99.82.83

1.111.08.58

1.231.451.371.381.10

1.141.211.021.041.061.131.15.97

1.01.77

1.191.00.93

1.051.041.16.97.54

1.241.051.18.92

1.31.83

1.141.08.74

1.171.26

1.061.061.15

Complexity

M

4.683.504.082.583.921.822.151.053.722.253.322.48

3.422.281.551.323.303.321.204.303.682.854.753.652.622.003.853.253.102.102.451.682.751.821.381.952.422.823.954.252.02

2.883.752.48

SD

.61

.10

.88

.70

.82

.67

.61

.31

.77

.89

.79

.74

.10

.81

.59

.47

.98

.93

.40

.84

.90

.79

.49

.82

.66

.59

.11

.73

.66

.66

.70

.79

.86

.80

.76

.67

.80

.74

.10

.77

.76

.68

.73

.90

K-F

1114

1569

9414001

12

—110104

2924

—1757

127110

18295

311

193137615237041

24434

210

131

18

A-A

4.832.593.694.88—

1.91—

3.97———

4.38

—1,342.031.90———

3.122.36—

2.28—

2.36—

2.45—

4.361.83

——

——

——

2.97—

2.312.97—

2.06——

(table continued)


Table Bl (continued)

45.46.47.48.49.50.51.52.53.54.55.56.57.58.59.60.61.62.63.64.65.66.67.68.69.70.

71.72.73.74.75.76.77.78.79.80.81.

82.83.84.85.86.

87.88.89.90.91.92.93.94.95.96.

Concept

CannonCapCarCarrotCatCaterpillarCeleryChainChairCherryChickenChiselChurchCigarCigaretteClockClothespinCloudClownCoatCombCornCouchCowCrownCup

DeerDeskDogDollDonkeyDoorDoorknobDressDresserDrumDuck

EagleEarElephantEnvelopeEye

FenceFingerFishFlagFlowerFluteFlyFootFootballFootball helmet

Name

H

.49

.591.08.00.00.96.83.16.00.52

1.352.33

.44

.00

.16

.16

.83

.17

.28

.95

.44

.88

.92

.44

.00

.44

1.44.32.00

1.42.87.16.38.00

2.55.00.28

1.14.28.00.16.16

.941.37.00.32.48.61

1.15.28.00.95

agree

%

908681

100100

797698

10083673393

10098988195957993816793

10093

7695

10071869890

100369895

7695

1009898

7471

1009593887695

10062

Image

M

3.522.683.104.503.782.383.754.463.224.523.623.152.982.754.652.203.722.853.252.593.784.083,053.922.853.65

3.723.183.052.283.483.803.902.303.223.713.85

3.494.263.854.704.15

3.804.603.583.223.253.413.224.424.184.38

agree

SD

.971.321.22.67.91

1.231.14.84

1.28.81

1.281.221.31.92.61.90

1.361.30.89

1.32.85.85

1.05.90.79

1.35

1.051.391.261.071.00.87

1.001.08.96

1.05.94

1.26.93.99.64.88

1,44.66

1.051.191.011.301.33.86.92.76

Familiarity

M

1.523.124.703.554.221.723.402.824.583.382.422.463.382.353.654.382.803.822.603.884.523.504.402.421.524.40

2.224.324.602.921.884.684.253.624.522.602.75

2.424.502,354.124.88

3.024.783.282.903.882.453.024.783.553.15

SD

.631.12.60.97.88.81

1.111.00.86

1.181.091.241.341.261.41.99

1.471.191.161.19.87

1.05.74

1.20.81.83

1.21.90.70

1.14.87.79.92

1.46.77

1.161.11

1.30.70

1.04.93.40

1.06.79

1.221.281.191.221.06.69

1.241.24

Complexity

M

3.922.184.052.953.253.584.252.552.051.603.483.123.283.582.252.682.822.124.502.552.383.582.283.854.251.78

3.553.053.384.123.353.222.682.652.952.883.32

4.182.684.121.423.48

2.552.303.751.883.254.154.102.182.282.98

SD

.82

.74

.95

.77

.94

.10

.86

.97

.70

.62

.90

.75

.11

.97

.77

.88

.92

.87

.81

.67

.83

.86

.84

.96

.77

.52

.77

.84

.73

.93

.69

.69

.61

.65

.89

.75

.82

.74

.82

.78

.591.10

1.00.95

1.02.46.94.85.92.89.71.69

K-F

727

2741

2314

50666

374

3481025200

283

436

3412291945

136575101

3123

671

119

529

721

122

30403516231

337036

—

A-A

——

——1.363.90—

—1.86——

—2.624.093.622.363.31

———

—2.94

—1.90

—1.66

——

1.551.55—

1.97

——

—2.48—

—1.822.853.932.00

——

2.61—

2.15——— •——

(table continued)



97.98.99.

100.101.

102.103.104.IDS.106.107.108.109.110.111.112.

113.114.115.116.117.118.119.120.121.122.

123.124.

125.

126.127.128.129.130.

131.132.133.134.135.136.137.138.139.140.141.142.143.

144.145.

Concept

ForkFoxFrench hornFrogFrying pan

Garbage canGiraffeGlassGlassesGloveGoatGorillaGrapesGrasshopperGuitarGun

HairHammerHandHangerHarpHatHeartHelicopterHorseHouse

IronIroning board

Jacket

KangarooKettleKeyKiteKnife

LadderLampLeafLegLemonLeopardLettuceLight bulbLight switchLionLipsLobsterLock

MittenMonkey

Name

H

.001.271.67.00

1.18

.76

.32

.161.07.16.77.79.38

1.47.16

1.09

.64

.00

.44

.74

.00

.16

.00

.32

.00

.32

.32

.56

.95

.001.66.00.00.60

.16

.44

.531.11.00

1.071.14.68.92.37.44.38.53

.96

.32

agree

%

1007457

10060

8895986498867690719874

9010093869398

10095

10095

9583

81

10040

10010090

98939081

1007674866793939088

7695

Image agree

M

4.153.493.733.603.92

4.524.484.403.813.653.463.584.313.554.203.82

2.714.104.304.734.283.654.493.424.202.65

4.084.40

2.22

4.303.314.584.103.25

3.753.263.883.644.353.683.054.424.623.884.103.623.51

3.823.12

SD

.851.201.361.02.93

.74

.811.00.94

1.201.261.07.79

1.301.211.05

1.181.02.90.55

1.061.22.98.97.81

1.11

.76

.77

.91

.751.11.74

1.001.32

1.14.90

1.121.05.94

1.031.20.83.62

1.03.94

1.351.40

.961.05

Familiarity

M

4.781.952.002.484.15

4.081.804.784.003.381.922.053.652.423.582.68

4.593.484.824.521.883.183.722.553.554.38

3.653.50

4.00

1.923.804.852.484.45

3.354.204.304.653.251.923.424.184.582.004.502.583.18

3.102.58

SD

.47

.841.051.05.96

1.10.95.52

1.301.061.061.181.041.071.091.19

.741.16.67.67

1.081.001.161.121.141.04

1.081.07

1.14

1.151.17.42

1.14.84

1.15.95.75.82

1.22.93

1.24.80.63

1.07.81

1.241.18

1.22.97

Complexity

M

2.624.024.303.422.05

2.584.651.822.853.023.183.623.004.404.003.52

2.882.602.981.204.052.351.003.803.823.90

3.252.05

3.25

' 3.982.401.922.851.92

2.321.852.522.551.854.283.482.752.524.301.854.482.22

2.353.90

SD

.94

.85

.871.05.67

.74

.73

.74

.85

.76

.77

.86

.92

.80

.92

.81

.78

.70

.91

.56

.81

.79

.00

.95

.70

.94

.89

.63

.80

.88

.74

.76

.69

.68

.61

.61

.77

.84

.69

.81

.92

.94

.77

.87

.88

.81

.69

.69

.70

K-F

1413

—1

— •

—0

992996070

19118

1489

43101

56173

1117591

43—

33

03

881

76

191812581800

——1769

123

09

A-A

2.24. ——

3.15—

—4.07——

3.12————

5.41—

—3.55——————

2.67—

3.88—

—

4.55—3.38

3.722.70

3.122.722.61—

3.064.18———

2.82—

5.28—

3.06—

(table continued)



146.147.148.149.150.

151.152.153.154.155.156.157.158.159.160.161.162.163.164.165.166.167.168.169.170.171.172.173.174.175.176.177.178.179.180.181.

182.183.184.185.186.187.188.189.190.191.192.

193.194.195.196.197.

Concept

MoonMotorcycleMountainMouseMushroom

NailNail fileNecklaceNeedleNoseNutOnionOrangeOstrichOwlPaintbrushPantsPeachPeacockPeanutPearPenPencilPenguinPepperPianoPigPineapplePipePitcherPliersPlugPocketbookPotPotatoPumpkin

RabbitRaccoonRecord playerRefrigeratorRhinocerosRingRocking chairRoller skateRolling pinRoosterRuler

SailboatSaltshakerSandwichSawScissors

Name

H

1.68.32.60.75.00

.161.041.88.86.16.97.00.53.35.00

1.06.53

1.19.81.37.00.32.00.38

1.07.70.60.00.16.54.38.29

1.72.86.34.00

.00

.581.73.44.56.16.53

1.00.94

1.21.16

.37

.96

.00

.16

.16

agree

%

6295907998

986760819864958186

1007488747993

10095

10090678190

10098888888

•57819098

10079509383989052717698

9383

1009898

Image agree

M

3.153.643.524.223.78

4.733.563.324.423.623.623.904.003.324.102.923.603.283.644.304.623.224.403.223.644.023.624.604.263.624.224.083.053.563.974.18

4.203.083.353.853.843.084.123.484.444.083.98

3.254.003.554.554.40

SD

1.581.051.12.91

1.11

.621.261.491.141.181.65.80

1.071.03

.921.59.92

1.28.97.98.62

1.04.80

1.151.281.061.04.62

1.00.84

1.151.02.92.98

1.141.18

.811.081.221.13.93.96.95

1.36.96.90

1.04

.991.12.97.77.83

Familiarity

M

3.983.252.702.452.88

3.283.152.703.404.522.553.323.341.522.222.784.552.902.053.003.554.784.421.702.923.422.182.952.903.503.384.183.954.223.463.08

2.952.204.404.681.523.483.252.252.222.223.58

2.924.184.452.923.98

SD

1.011.091.191.021.23

1.201.391.311.14.87

1.281.311.26.67

1.061.24.86

1.021.051.021.14.72

1.00.93

1.291.48.97

1.301.14.92

1.13.77

1.28.96

1.171.35

1.071.23.86.65.89

1.281.301.111.081.08.95

1.17.92.97

1.19.99

Complexity

M

1.024.782.803.283.12

1.803.181.781.551.602.302.852.123.704.222.582.222.554.752.821.153.152.322.822.484.583.004.351.881.852.202.252.702.222.202.60

3.284.403.322.204.152.553.584.081.524.121.85

3.583.003.422.252.15

SD

.16

.471.05.87.71

.681.00.88.74.92.56.96.71.81.72.95.70.81.43.95.36.94.91.70.95.77.81

1.01.71.57.60.70.78.69

1.10.70

.84

.83

.93

.60

.85.80.92.93.50.90.94

.92

.92

.86

.62

.65

K-F

600

3310

2

6—

3156015152302193266

18340

133889

20211

233

28152

111

—233

47———

33

10

10352

1

A-A

——

3.21——

3.58———

—————

———

2.795.18————

5.12——

2.94—

4.074.07————

2.673.69

2.61———

—— •————

—————

(table continued)



198.199.200.201.202.203.204.205.206.207.208.209.210.211.212.213.214.215.216.217.218.219.220.221.222.223.224.225.

226.227.228.229.230.231.232.233.234.235.236.237.238.239.240.241.242.243.244.

245.

246.247.248.

249.250.

Concept

ScrewScrewdriverSea horseSealSheepShirtShoeSkirtSkunkSledSnailSnakeSnowmanSockSpiderSpinning wheelSpool of threadSpoonSquirrelStarStoolStoveStrawberrySuitcaseSunSwanSweaterSwing

TableTelephoneTelevisionTennis racketThimbleThumbTieTigerToasterToeTomatoToothbrushTopTraffic lightTrainTreeTruckTrumpetTurtle

Umbrella

VaseVestViolin

WagonWatch

Name

H

.33

.00

.34

.61

.95

.00

.28

.16

.16

.00

.51

.16

.00

.00

.612.041.54.16.17.00.16

1.12.17

1.01.00.64.98.17

.32

.591.46.62.63.16.89.33.00

1.57.80.16.68

1.40.74.00.53

1.10.32

.00

.32

.16

.72

.92

.45

agree

%

9398888867

100959898988698

1001008850559893

10098769079

100888395

9586528683986993

100558898866786

100907995

100

959886

7990

Image

M

3.674.303.583.183.003.863.023.283.404.493.333.544.003.722.953.103.804.104.424.414.124.103.982.984.223.692.784.15

3.424.284.004.624.264.484.053.823.924.184.054.403.464.083.203.522.802.894.12

3.92

2.723.704.18

3.563.18

agree

SD

.89

.641.221.061.11.98

1.261.101.09.81

1.181.01.95

1.001.161.321.441.11.89

1.101.081.001.041.171.08

.721.11.92

1.361.16.82.58.93.63.94

1.14.79.83

1.12.74

1.05.98

1.381.001.101.42.95

.90

1.021.101.05

1.631.07

Familiarity

M

3.203.421.501.621.854.564.623.642.302.801.851.903.154.522.281.183.124.503.823.353.084.653.203.654.901.974.483.02

4.354.804.823.622.484.723.802.104.084.483.784.621.884.554.154.684.022.602.40

3.95

2.783.482.68

2.504.58

SD

1.001.14.89.73.82.70.70

1.531.171.031.061.041.04.84

1.10.54

1.14.89.89

1.331.13.65

1.29.91.30.83.74

1.24

.88

.51

.381.301.12.74

1.03.92.90.81

1.06.73.98.80.88.61.91

1.261.14

.92

1.261.051.21

1.22.73

Complexity

M

3.252.354.502.903.803.083.381.404.723.053.404.522.521.623.684.253.182.023.751.052.324.023.383.601.203.052.902.72

1.723.523.223.253.352.382.904.622.781.981.982.422.653.454.323.702.753.583.62

3.00

3.152.604.10

3.353.40

SD

.99

.73

.71

.74

.75

.79

.86

.58

.74

.84

.80

.81

.59

.62

.85

.92

.97

.82

.97

.22

.72

.94

.91

.86

.46

.80

.77

.97

.77

.97

.96

.94

.82

.97

.80

.80

.85

.82

.57

.77

.82

.84

.88

.81

.86

.92

.89

1.05

.66

.74

.86

.911.04

K-F

210

—1723271421001

44042

——

61

258

150

20112

31424

1987650— .

11023

70946

204—825957

78

8

44

11

5581

A-A

4.45————— .

1.94

——

—5.103.52—————

1.97— •—

—2.723.61—

————

2.45—

2.62————————————

2.03——

2.97

4.09

4.30

——

3.03—

(table continued)



Concept

260. Zebra

Name agree Image agree Familiarity Complexity

% M SD M SD M SD

.00 98 4.05 .74 1.60 .83 4.55 .70

K-F A-A

251.252.253.254.255.256.257.258.259.

Watering canWatermelonWellWheelWhistleWindmillWindowWineglassWrench

2.03.55.60.33.00.16.32

1.43.89

55869093

10098955076

4.082.854.183.484.553.353.253.312.51

.981.31.86

1.36.67

1.08.83

1.351.18

2.723.051.452.222.451.804.404.022.72

1.501.09.70

1.04.92

1.00.86

1.111.28

2.782.283.822.422.554.623.181.852.02

.79

.92

.74

.83

.84

.76

.86

.48

.79

—1

89756

41

11900

—————

5.482.28——

1 —

Appendix C

Shown here are all the concepts for which one or more naming, imaging, or identificationfailure occurred, a different object was imaged, or more than one name was given. Failuresin the naming task are listed as DKO (don't know object), DKN (don't know name) andTOT (tip of the tongue). Identification failures in the familiarity rating task are listed asDKO (FAM). Imaging failures in the image agreement task are listed as NI (no image) andimaging a different object as DO (different object). All nondominant names given for eachconcept are listed accompanied by their frequencies. For names given by a single subject,only the name is listed without the frequency.

Table ClNorms

DKOConcept DKO DKN TOT (FAM) NI DO Nondominant names

1. Accordion 0 1 3 0 0 02 . Airplane 0 0 0 0 1 0

3 . Alligator 0 0 1 0 0 04 . Anchor 0 2 0 0 0 05 . A n t 1 3 0 0 1 16 . Apple 0 0 0 0 0 07 . A r m 0 0 0 0 0 08 . Arrow 0 0 0 1 0 79 . Artichoke 3 6 3 5 1 5 1

1 0 . Ashtray 0 0 0 1 0 01 1 . Asparagus 1 3 1 2 3 1

1 2 . A x e 0 0 0 0 0 0

1 3 . Baby carriage 0 0 1 0 0 01 4 . Ball 0 0 0 0 1 01 5 . Balloon 0 0 0 0 0 11 7 . Barn 0 0 0 0 0 0

OrganPlane 8, jet 4, jet plane 2,

jet airplane 2, airplane-747Crocodile 3, lizardShip's anchorInsect 2, spider, bugFruitLeft arm 3, handOne-way signAvocado 4, pineapple, bud, brussel,

sprout, squash

Asparagus spear 3, branch of willows,twig, pine tree, cauliflower, branch

Hatchet 3, hammer

Carriage 19Beach ball 2, top

Barn and silo 9, farm 2, barn house,farmhouse

(table continued)


Table Cl (continued)

Concept

18. Barrel19. Baseball bat20. Basket

21. Bear23. Bee

24. Beetle

25. Bell26. Belt27. Bicycle28. Bird29. Blouse

30. Book31. Boot32. Bottle33. Bow34. Bowl35. Box

36. Bread

38. Brush

41. Button

42. Cake

43. Camel45. Cannon46. Cap47. Car

50. Caterpillar51. Celery52. Chain54. Cherry55. Chicken56. Chisel

57. Church59. Cigarette60. Clock61. Clothespin62. Cloud63. Clown64. Coat65. Comb66. Corn67. Couch68. Cow70. Cup

DKO

000

00

0

00000

000000

0

0

0

0

0000

100104

000000000000

DKN

000

02

2

00000

000000

0

0

0

0

0100

130205

000010000000

TOT

000

00

1

00000

000010

0

0

0

0

2000

010102

000200000000

DKO(FAM)

000

00

0

00000

000000

0

0

0

0

0000

000001

000070000000

NI

000

00

1

00011

100000

0

0

0

0

0010

000005

000000100000

DO

101

00

0

11000

010

1320

0

0

0

0

0020

101001

000000000000

Nondominant names

Bat 20Picnic basket 2, lunch basket,

wicker basketPolar bear 5Fly 8, insect 3, bug 2, bumble bee,

mosquitoInsect 4, bug 4, cockroach 4, roach 3,

cricket, ant, dico

—CollarBike 5Sparrow 3, chickadee, baby chickShirt 16, jacket 4, sweater, coat,

ladies jacket, stylish shirt

—Shoe 3, rubber boot, half-bootWine bottle 2Ribbon 6, bowtie 3, tie, ribbon bowDishCube, shoe box, index card box,

cardboard box, index file boxLoaf of bread 5, slice of bread,

loaf of bread and sliceHair brush 4, shoe brush 2, lint

brushWheel

Layer cake 5, piece of cake,three-layer cake

—War gun, gun, cannon (civil war)Hat6Lincoln 4, automobile, motor car,

luxury car, luxury automobileCentipede 4, snail, worm, inchwormLettuce 4, staff of celery, celery stalkChain linksPeach, grape, plumHen 10, turkey 2, rooster, birdScrewdriver 8, tool 2, scraper 2,

wedge, awl, wood chisel, file,knife

Chapel 2, church houseBurning cigaretteMantle clockClothes clip 3, clip 2, clothes holderBushesClown face 2Overcoat 6, jacket 3Haircomb 2, pocket combEar of corn 5, corn on a cob 3Sofa 14Bull 2, female cowCoffee cup 2, teacup

(table continued)



Concept

71. Deer

72. Desk74. Doll

75. Donkey76. Door77. Doorknob79. Dresser

80. Drum81. Duck

82. Eagle83. Ear85. Envelope86. Eye

87. Fence88. Finger

90. Flag91. Flower92. Flute93. Fly94. Foot96. Football helmet98. Fox99. French horn

101. Frying pan

102. Garbage can

103. Giraffe104. Glass105. Glasses106. Glove107. Goat108. Gorilla109. Grapes110. Grasshopper

111. Guitar112. Gun

113. Hair

115. Hand116. Hanger117. Harp118. Hat119. Heart120. Helicopter122. House

DKO

0

00

0000

00

0000

00

00100000

0

0

00000000

00

0

0000000

DKN

0

00

0000

00

1000

00

00020003

0

0

00000001

00

0

0000000

TOT

0

00

0010

10

0000

00

00000102

0

0

00001010

00

0

0030000

DKO(FAM)

0

01

0000

00

0000

00

00000000

0

0

00000000

00

1

0000000

NI

0

00

0000

10

1000

00

00000004

0

0

01n01000

00

1

0010000

DO

0

10

0000

10

0100

10

04100013

0

0

02400000

00

6

0300300

Nondominant names

Antelope 3, reindeer 2, elk 2, ram,buck, stag

Bench, office deskBaby 5, little girl 4, female doll,

child, baby dollMule 2, pony 2, burro, jackassWooden doorKnob 3Bureau 7, chest of drawers 6, chest 6,

drawers 4, dresser drawers 2,bureau drawers, desk

—Goose 2

Bird 4, hawk 3, bald eagle, parrotRight ear 2Sealed envelopeEyeball

Picket fence 10, gateIndex finger 7, right index finger 2,

pointing finger, pointer, forefingerFlag and staff, flag poleRose, marigold, daisyClarinet 2, coronet, windpipeBug 3, bee 2, insect, wasp, spiderLeft foot 2Helmet 15Wolf 6, dog 2, coyote 2, hound dogTuba 4, horn 4, trumpet 3,

trombone, coronetPan 15, fry pan 2

Trash can 2, waste can, garbagepail, garbage

Ostrich, zebraCupEve glasses 14, wire-rim glassesRight gloveBilly goat 2, mule, horse, donkeyApe 10Bunch of grapes 3Bug 4, cricket 2, spider 2, bee,

hornet, cicadaAcoustic guitarPistol 6, revolver 5

Straight hair, women's hair style,hair style, head of hair

Right hand 2, fingersClothes hanger 3, coat hanger 3—Felt hat—Plane, copterCottage, private house

(table continued)



Concept

123. Iron124. Ironing board

125. Jacket

127. Kettle130. Knife

131. Ladder132. Lamp133. Leaf134. Leg

136. Leopard137. Lettuce

138. Light bulb139. Light switch140. Lion141. Lips142. Lobster143. Lock

144. Mitten145. Monkey146. Moon

147. Motorcycle148. Mountain149. Mouse

150. Mushroom151. Nail152. Nail file153. Necklace

154. Needle155. Nose156. Nut

157. Onion158. Orange159. Ostrich

161. Paintbrush162. Pants163. Peach164. Peacock165. Peanut167. Pen169. Penguin170. Pepper

171. Piano172. Pig174. Pipe

DKO

00

0

00

0000

02

000000

000

000

0000

000

140

00200001000

DKN

01

0

00

0000

00

000010

000

000

0000

003

002

00000012

000

TOT

02

0

00

0000

20

000000

000

000

1000

002

112

00140001

000

DKO(FAM)

00

0

00

0000

00

000000

000

000

0000

000

020

00200001000

NI

10

0

00

0000

00

000000

100

100

0010

001002

00010000

000

DO

00

0

10

000100

000005

001

000

0700

20

31

000

10000007

005

Nondominant names

Cloth iron, steam ironIron board 3, ironing table

Shirt 5, coat 2, sport jacket

Tea kettle 15, teapot 9, potEating knife 2, butter knife, table

knife

Part of a ladderTable lamp 2, lamp & shadeMaple leaf 3, oak leafKnee 3, right leg 2, knee to foot,

foot, calfTiger 4, panther 2, jaguar, checqwaCabbage 5, cauliflower 2, celery,

head of lettuceBulb 5, incandescent light bulbSwitch 14Male lion 3Mouth 2, pair of lipsCrab 3Padlock 5

Glove 8, left mitten 2Chimp, chimpanzeeQuarter moon 6, crescent moon 4,

half moon 4, crescent 2Motor bike, bikeMountain peak 2, snowcap, hillRat 9

SpikeFile 13, fingernail filePearl necklace 5, beads 5, pearls 4,

chain, pearl chain, string of pearlsPin 6, sewing needle 2Left nostrilBolt 9, hexagonal nut

—Ball, grapefruit, fruitStork, turkey

Brush 9, ink pen, mk penSlacks 5Apple 2, pear 2, plum 2, fruit, orangeTurkey 2, rooster, ostrich, birdNut 3Writing pen, ball point penPelican 3Green pepper 8, artichoke, bell

pepperGrand piano 8Hog 2, rhinoceros, bullPipe (tobacco)

(table continued)



Concept

175. Pitcher176. Pliers177. Plug178. Pocketbook

179. Pot180. Potato181. Pumpkin

183. Raccoon184. Record player

185. Refrigerator186. Rhinoceros187. Ring188. Rocking chair189. Roller skate190. Rolling pin191. Rooster192. Ruler

193. Sailboat194. Salt shaker

196. Saw197. Scissors198. Screw199. Screwdriver200. Sea horse201. Seal202. Sheep204. Shoe205. Skirt206. Skunk207. Sled208. Snail209. Snake212. Spider213. Spinning wheel

214. Spool of thread

215. Spoon216. Squirrel217. Star218. Stool219. Stove220. Strawberry221. Suitcase

223. Swan224. Sweater

225. Swing

DKO

0000

020

10

00000000

00

000000000001000

0

0000010

00

0

DKN

1110

000

10

01000300

00

000030300001015

0

0000010

00

1

TOT

0120

001

31

02000100

00

001101100011003

0

0200011

00

0

DKO(FAM)

0000

010

00

00000000

00

000000001000000

0

0000000

00

0

NI

0000

010

20

02000100

00

000000100211005

0

0000000

10

1

DO

6030

400

00

00000300

00

001001000000105

0

0010000

00

0

Nondominant names

Jug 3, water pitcherWrench 3Electric plug 2Handbag 7, purse 7, bag 3,

shoulder bagPan 6, saucepan 2Peanut, acorn—

Fox 3, badgerPhonograph 12, turntable 5,

stereo 2, phonoIce box 2, refrigidaireRhino 3, hippopotamusPearl ringChair 3, rockerSkate 20Roller 5, dough roller 3Chicken 5, hen 2, turkey 2, cock12-inch ruler

Boat 3Salt or pepper shaker 3, shaker 2,

salt container, pepper shakerHand sawSewing scissorsScrew-round head wood

—Fish, dragonWalrus 2, otter, sea lionLamb 9, bullRight shoe 2DressRaccoon

—Slug, shell, snail shellCobraInsect 2, praying mantis, bugLoom 3, sewing machine 3, spindle

2, wheel to make clothes,knitting wheel, knittingmill, wool, spinwheel

Thread 11, spool 7, spool of sewthread

TeaspoonChipmunk

—ChairOven 5, range 4, gas stoveRaspberryLuggage 5, valise, piece of luggage,

canvas luggageGoose 3, duck 2Pullover 2, shirt 2, sweat shirt 2,

pull over sweaterSwinging chair

(table continued)



Concept

226. Table227. Telephone228. Television229. Tennis racket230. Thimble

231. Thumb232. Tie233. Tiger235. Toe236. Tomato237. Toothbrush238. Top239. Traffic light

240. Train242. Truck243. Trumpet244. Turtle

246. Vase247. Vest248. Violin249. Wagon250. Watch251. Watering can

252. Watermelon253. Well

254. Wheel256. Windmill257. Window258. Wineglass259. Wrench

260. Zebra

DKO

00000

00000000

0000

000000

20

00000

0

DKN

00000

00000000

0000

000101

00

00005

1

TOT

00013

00100000

0000

000104

00

10000

0

DKO(FAM)

00000

00000000

0000

000000

00

00000

0

NI

00000

00000020

0000

000002

00

00002

0

DO

00101

02000050

0000

000600

10

00003

0

Nondominant names

Desk, benchPhone 6TV 12, television set 8Racket 4, paddleNimble 2, thumb nimble, thumb

capFingerNecktie 13Leopard, lionBig toe 12, toes 5, right big toe 2Pepper, radish, onion, peach, fruitBrushSpinning top 5, draddleStop light 10, street light, signal

light, crossing lights, lightLocomotive 3, train engine 3Tractor trailer 3, trailerHorn 5, bugle 2, tuba, coronetTortoise, box turtle

Flowered vaseVestcoatBase 4, cello 2Cart 3, wheelbarrow 3, carriageWrist watch 4Sprinkler 5, watering pot 2, pitcher,

water planter, flower waterer,water pitcher, water can, waterbasket, pot

Watermelon slice 3, melonWater well 2, waterbucket,

wishing wellWagon wheel, spoked wheelWindbrakerWindow-double hung, closed windowGlass 15, goblet 6Pliers, belt turner, spanouke,

open end wrench, crescent wrench

—

Received June 18, 1979Revision received August 24, 1979 •

Documents

snodgrass (1980) a standardized set of 260 pictures. norms for name agreement, image agreement, familiarity, and visual complexity