Subjective Experience of Word Production Difficulties in Aphasia: a Metaphor Analysis of Autobiographical Accounts

ABSTRACT Background and aims The subjective experience of neurological symptoms provides useful information for assessment, intervention and care. However, research in the subjective experience of aphasia is limited. Word production difficulties are universal to aphasia, and interdisciplinary research has produced sophisticated models of the multiple stages and processes involved. Critically, this word-production research does not incorporate the subjective experience of symptoms. We carried out a metaphor-led discourse analysis on autobiographical accounts written by people with aphasia, to determine whether subjective descriptions of word finding difficulties are consistent with the stages and processes of psycholinguistic models. Method Metaphor-led discourse analysis was used to identify, code and interpret metaphorical expressions describing word production difficulties in 12 English-language autobiographical accounts written by people with aphasia. These expressions were then analysed to determine the systematic metaphors (i.e., the related concepts which are used consistently to describe a particular topic). Two distinct types of systematic metaphor emerged in the analysis: conventional systematic metaphors frequently recurring throughout all or most of the accounts; novel systematic metaphors used in one or two extended passages in an overlapping subset of the accounts. Results and discussion 4020 metaphorical expressions described word production, predominantly using conventional metaphors about communication and cognition. The conventional metaphor WORD-PRODUCTION AS MOVING OBJECTS OUT OF A CONTAINER was the most prevalent, with elaborations and variations allowing mapping of different symptoms. Other conventional metaphors included: WORD PRODUCTION AS A JOURNEY/HUNTING/HERDING THROUGH A LANDSCAPE, allowing description of effortful or partial retrieval, neuroplastic recovery, and internal strategies; APHASIA AS BODILY IMPAIRMENT, which described various symptoms in terms of different body parts, including self-monitoring difficulties; and APHASIA AS FRAGMENTATION AND PERSONIFICATION OF SELF and SELF AS MACHINE/COMPUTER to describe a disrupted sense of agency and attention. Novel systematic metaphors were used to describe certain symptoms: APHASIA AS SILENCE and APHASIA AS SPIRITUAL EXPERIENCE were used to describe a lack of ‘inner speech’, and APHASIA AS A DISMEMBERED TREE to describe problems making semantic associations. Conclusions and implications This research demonstrates the many consistencies of subjective descriptions of word production difficulties in aphasia with theoretical models, but also shows that some subjectively salient symptoms, in particular attentional and self-monitoring difficulties, and a lack of inner speech, are not captured by all theoretical models. Careful attention to the way that people with aphasia describe their symptoms can provide a valuable source of information to be integrated with objective measures.

and APHASIA AS A DISMEMBERED TREE to describe problems making semantic associations.Conclusions and implications: This research demonstrates the many consistencies of subjective descriptions of word production difficulties in aphasia with theoretical models, but also shows that some subjectively salient symptoms, in particular attentional and self-monitoring difficulties, and a lack of inner speech, are not captured by all theoretical models.Careful attention to the way that people with aphasia describe their symptoms can provide a valuable source of information to be integrated with objective measures.

Background
Clinically, the subjective experience of neurological symptoms provides useful information for assessment, intervention and care, in some cases providing the main source of information for diagnosis.This can include descriptions of common experiences such as pain or fatigue, and of more unusual symptoms such as perceptual distortions.Subjective symptoms can differentiate behaviourally similar disorders, for example, self-reported mood can distinguish between pseudobulbar affect and depression as the cause of crying episodes following brain injury (Engelman et al., 2014).Emerging research suggests that subjective experience sheds light on the neuropsychological profile of complex and heterogenous syndromes such as dementia (Zwijsen et al., 2016).In aphasiology there is limited research on subjective experience, likely primarily due to the difficulty of describing symptoms when communication is impaired.This study demonstrates that despite the difficulties, there is valuable diagnostic information to be gained from subjective descriptions of aphasia.
The potential contribution of subjective experience to understanding the underlying dynamics of word production can be illustrated with the example of a semantic fluency task.After such a task (e.g., listing animals) a client might say: "I can hear it in my head, but I can't say it," which suggests successful phonological activation but difficulty with articulation; or "another word gets in the way," perhaps suggesting a selection difficulty; or "the word keeps slipping away" suggesting working memory problems.These subjective reports appear to be indicative of difficulties at particular stages (e.g., semantic, phonological, articulatory) and also of how processing is affected (e.g., activation, selection, retention).This information is not available in the behavioural measures of this task such as the number of correct words.Thus there are missed opportunities when such insights into language impairment are not routinely captured.
This research focuses on word production difficulties, as these are a universal feature of aphasia, regardless of subtype or severity.Extensive psycholinguistic and neuropsychological research has resulted in sophisticated models of word production, which explain the different presentations of word production difficulties in terms of impairment to semantic, phonological or articulatory stages.Models differ on the degree of modularity of these stages, and on how integral self-monitoring through receptive language networks, and cognitive processes such as attention, are to word production (e.g., compare Nozari et al., 2011 andRoelofs, 2014).Critically, while discussions about the relative merits of such models draw on a range of methodologies, they omit much of the information available in subjective experiences of word finding difficulties.This is therefore an underexplored source of valuable information on hard-to-observe underlying processes.

Subjective experience of aphasic symptoms
The subjective experience of aphasia, while currently often neglected, was foundational in the establishment of scientific aphasiology.Lordat's (1843) first-hand account of transient word production difficulties was much discussed by early theorists such as Wernicke (1874) (see Bay, 1969).In the intervening two centuries, autobiographical accounts of aphasia have increased in number, length, the severity of symptoms described, and demographic diversity, due to better medical treatment and widespread access to education and the means to write and publish.However, as aphasiology has become more specialised, the insights available in first-hand accounts have become detached from theoretical research.
Luria is a notable exception, as he routinely included patients' descriptions of symptoms in Traumatic Aphasia (e.g., "Everything immediately flew away" . . ."the words fell to pieces," p128, 1970), and edited and contextualised his patient Zasetsky's autobiographical journal in The Man with a Shattered World (1972).In this work he combined subjective descriptions with behavioural testing to prise apart the stages of processing, for example "We have already pointed out that he could not recall a word immediately but had to search actively for it, often finding that other words occurred to him instead . . .How, then, was he to select the right word when his memory was cluttered with words . . .? " (p107, 1972).Yet this close attention to the subjective experience has not had the same influence as the objective aspects of Luria's work which it informed (Sacks, 2014).
There have been sporadic attempts to integrate the subjective perspective with finegrained pyscholinguistic theory.Rolnick and Hoops (1969) suggest that subjective symptoms can provide "insight into the underlying dynamics of the disorder."Their analysis of interview data from six individuals with aphasia found descriptions of processes such as self-monitoring, the effect of speaking rate on word finding, and difficulty with certain syntactic categories.More recently, Fama et al. (2022) carried out a thematic analysis of interview data from fifty-three people with aphasia.Much of the analysis concerns the impact-level factors, but under the theme 'mechanism' there is some exploration of how word production were described at the level of impairment (World Health Organisation, 2001).For example, subjective descriptions of problems with word 'memory' as likely referring to a 'failure of lexical access/retrieval.'While this work shows the greater rigor made possible by modern qualitative methods and analysis of a larger amount of data, it does not analyse descriptions of impairment systematically and at the same level of detail as Rolnick and Hoops (1969).
Three other recent studies interpret subjective reports in terms of specific fine-grained psycholinguistic theory (Ardila & Rubio-Bruno, 2018;Morin, 2009;Skipper, 2022).These studies do not attempt methodological rigour, but select short extracts of subjective accounts which provide illustration or anecdotal evidence for the authors' theoretical interests.Taking a different approach there are studies which demonstrate that subjectively reported symptoms in aphasia can be confirmed to correlate with objective measures, such as the occurrence of tip-of-the-tongue states (Goodglass et al., 1976), or intact inner speech (Fama & Turkeltaub, 2020).These studies show that subjective data can be integrated in a rigorous way, but with the limitation that participants are asked whether or not they agree with a statement which is predetermined by the researchers.Such an approach loses the fine-grained detail and the ability to capture novel insights which are available with a bottom-up approach.
Through the novel application of metaphor-led discourse analysis to subjective descriptions of impairment in aphasia, this study brings together the systematic and bottom-up application of formal qualitative methods with the granularity of detail found in psycholinguistic models.This method is ideally suited to the exploration of a large quantity of autobiographical writing from multiple authors, as it can accommodate the heterogeneity of experience represented in these accounts while still allowing meaningful comparisons to be made.Taking such an empirical linguistic approach lessens the impact of researchers' preconceptions and interests on the analysis (Cameron & Maslen, 2010).

Metaphor-led discourse analysis for exploring the subjective experience of aphasia
Metaphor-led discourse analysis is a rigorous approach to exploring subjective experience through qualitative analysis of discourse data such as interviews or written accounts (Cameron & Maslen, 2010).In healthcare, the most influential application of this method has been Semino's work on people's experience of cancer and its treatment (e.g., Semino et al., 2015).Emerging work demonstrates that this method can be applied not only to physical illnesses or to questions of impact, but to impairment-level exploration of symptoms of neurological and psychiatric conditions.To give a noteworthy example, metaphor analysis of clinical interview transcripts can differentiate epileptic from nonepileptic seizures, highlighting the potential for this approach to reveal the underlying mental processes behind similar observable behaviours by systematically capturing information in the subtleties of description (Plug et al., 2009).This information may be of a type which an experienced clinician is sensitive to, but which is neither transparent to introspection nor easily formalised, and therefore not readily available to report and teach.A review of various psychiatric conditions such as depression and schizophrenia established that complex and heterogenous alterations of cognitive and affective processes can be effectively explored through metaphor-led discourse analysis (Littlemore, 2019).These studies provide a firm grounding to extend the approach, to explore whether the details of the cognitive and linguistic processes affected in word production difficulties in aphasia are reflected in the metaphors used to describe the experience.
The use of metaphors to explore experience is based on an understanding of metaphors as more than a matter of rhetoric.According to Conceptual Metaphor Theory, metaphors play a central role in shaping our understanding of abstract and complex 'target domains' through the mapping of more familiar and concrete 'source domains' (Lakoff & Johnson, 1980).Through metaphorical mappings, we relate some aspects of the source domain to relevant or salient aspects of the target domain, giving us a structure for understanding it.Thus by analyzing the metaphorical language used to describe cognitive or linguistic processes, we can gain insight into the underlying patterns of thought and experience.To give an example which has been discussed in the metaphor literature, UNDERSTANDING AS SEEING 1 maps the source domain of vision onto the target domain of cognition, as in "I see what you mean."Close analysis of the use of this metaphor in discourse data shows that it is used specifically to describe the process of gaining an understanding of another person's perspective (Deignan & Cameron, 2009), demonstrating how empirical methods can reveal nuances unavailable to intuition.

The Current Study
Written autobiographies provide a unique source of discourse data on the subjective experience of symptoms, to which a metaphor analysis can be applied.In this study 12 autobiographies were selected, written by authors with a wide range of aphasia types and severity (from mild and transient with full recovery, to global aphasia with significant persisting difficulties), different aetiologies (ten stroke, one traumatic brain injury, one cancer), male (four) and female (eight), ranging in age from 27 to 79, and spanning a range of styles and genre, from factual and reflective self-published journals such as Broussard's (2016) Stroke Diary, to the creative memoir of a novelist, The Shadow Factory, (West, 2008), and Taylor's (2009) professionally published popular science best seller My Stroke of Insight.
Long form written accounts have advantages over interview data as the author does not need to make pragmatic compromises in their word choice because of the time pressures of an interview, nor is their choice of language as influenced by the clinical or research environment.As the method does not place undue weight on any single metaphorical expression, but looks for systematic application of similar metaphors, any distortion of the authors' words due to editorial processes should not affect the analysis.Metaphor analysis has been carried out on published first-hand accounts in other areas; Vidali (2010) includes autobiography, poetry, and academic writings produced by authors with disabilities in her discussion of embodied metaphor.Published works are considered as directly attributable to their authors, despite the ways in which the world of publishing deviates from that of carefully controlled research.The unknown contribution of support with editing and writing is offset by the inclusion of a range of types of texts produced and published in a wide variety of ways, including the inclusion in some of the accounts of extracts of unedited contemporaneous journal entries.The selection bias that may arise from a sample of individuals who are able to write and publish their own accounts is minimized through careful and constrained interpretation of the analysis.All of these accounts contain the detailed reports of years of intensive work on language recovery by intelligent and highly motivated people, some of whom have relevant prior expertise, such as Jill Bolte Taylor's background in neuroscience or Carol Maloney's experience working in education and dyslexia, or who apply expertise and analytical tools from other disciplines, such as Thomas Broussard's engineering-informed development of tools and systems for collecting information about his impairment and careful logical analysis of the information thus gained.Each account offers a depth of reflection on a particular case of aphasia which is available from no other source.
All metaphors used to describe word production in these accounts were identified following Cameron and Maslen (2010), see methods.The use of metaphorical language in describing aphasia symptoms may be in the form of a single word (such as "evaporated") or a longer phrase or sentence (such as "the fragments of my broken speech scattered around me").These linguistic units are termed the metaphor 'vehicles.'The vehicle terms were coded in an iterative procedure into 'vehicle groups' based on the semantic meaning of the source domain (e.g., 'evaporation,' or 'disintegration').Subsequently, through repeated examination of the sorted data, with consideration of the meaning in context of the coded expressions, systematic metaphors were identified, for example, WORD PRODUCTION DIFFICULTIES AS A LOSS OF SOLIDITY.
Through exploration of how these systematic metaphors are used in the autobiographical accounts, the commonalities and differences of the authors' symptoms can be explored, and insight gained into the range of experiences described (Littlemore, 2019).Consideration of the way in which these metaphors are used, with reference to the contextual information provided about the authors' symptoms, allows for suggestions to be made about whether different theoretical descriptions of impaired processing could account for the symptoms described.
Models of word production agree on the existence of different stages of processing, with broad agreement that distinctions can be made between semantic, phonological and motor levels.These levels may be differentially impaired.Models differ in various dimensions: the degree to which these levels are sequential and encapsulated modules or to which activation occurs between levels before each has selected a representation; whether self-monitoring is primarily within the domain of receptive language networks or part of motor planning and prediction; the role of cognitive processes such as attention in word production (e.g., compare Nozari et al., 2011vs. Roelofs, 2014).The current study offers a way of examining these differences through the descriptions of subjective experience of people with aphasia.To give an example, the 'evaporation' of a word appears to describe the successful retrieval of a phonological word form followed by difficulty retaining it in working memory, whereas the 'melting' or 'disintegration' of a word is used to describe successful activation followed by difficulty with motor planning or articulation (these metaphors are explored and illustrated in Results).This example shows the value of analysing a large amount of data in identifying the way that different symptoms may be conceptualised using different mappings available within a particular systematic metaphor.This systematicity means that this work can form the basis for future mixed methods work.
Research in metaphor has identified numerous conventional metaphors, which are those that a linguistic community uses to conceptualise a particular target domain.The conventionality or novelty of the systematic metaphors found in this analysis was considered, that is, whether the metaphors used conformed to the usual ways in which we conceptualise language use.This allows us to identify the aspects of the experiences described which are subjectively more unusual.These unusual experiences are perhaps harder for clinicians, researchers and family members to understand empathically.Attention to the metaphors used has the potential to facilitate understanding of such symptoms and to improve clinical rapport.
The aims of this study are 1) to explore which metaphors are used to describe the subjective experiences of word finding difficulties in written accounts by people with aphasia, 2) to consider whether these descriptions reflect the details of the cognitive and linguistic impairment as described by psycholinguistic theory, and 3) to consider whether particular processes or symptoms which are described are not accounted for in some models or theories.

Data Source
Search engines and library catalogues were used to find accounts written in English by a person with aphasia.Twelve autobiographical accounts of aphasia were included (eleven books, one short extract), these were: Stroke Diary (Broussard, 2016), My Stroke of Luck (Douglas, 2002), Until Further Notice I am Alive (Lubbock, 2012), Finding My Voice with Aphasia (Maloney, 2013), A Stitch in Time (Marks, 2017), A Mind of My Own (Mills, 2004), Without Utterance (Resch, 2012), Crossing the Void (Schultz, 2010), My Stroke of Insight (Taylor, 2009), The Shadow Factory (West, 2008), Aphasia, my world alone (Wulf, 1973), and a short extract by 'Maria,' from Jumbly Words, and Rights Where Wrongs Should Be: The Experience of Aphasia from the Inside (eds.Edelman & Greenwood, 1992).See Appendix A for a summary of background information and aphasia characteristics provided in the accounts analysed.

Metaphor identification, coding and analysis
A five-stage process of metaphor identification, coding and analysis was used, following Cameron et al.'s (2009) metaphor-led discourse dynamic method (see Figure 1).Metaphor identification and coding were carried out by the first author, a Speech and Language Therapist with clinical experience in aphasia, and three research assistants (Speech and Language Therapy students), each of whom worked with the same 3-5 books for each step of the analysis.Description and interpretation of systematic metaphors was carried out by the first author.We provide details in Appendix B of the measures undertaken to increase rigour, reduce bias and enhance replicability at each of the five steps of analysis.Researchers wishing to explore to the dataset are invited to contact the corresponding author for access.

Step 1. Familiarisation with Texts and Selection of Descriptions of Language Processing
First, all researchers familiarised themselves with the texts and identified all descriptions of the subjective experience of word-production.For each text a minimum of two researchers independently identified relevant sections, with discussion of any points of disagreement.An inclusive approach was taken to difficult cases, for example with the

5.Interpretation of Systematic Metaphors
Codified procedures, collaborative and iterative process, produces quantifiable data Subjective interpretation of single researcher Following Pragglejaz group (2007) Following Cameron & Maslen (2010) Informed by psycholinguistic literature Figure 1.Five stage process of metaphor identification, coding and interpretation inclusion of descriptions of speech, writing and inner speech, and descriptions of the general state of having a language impairment, as well as descriptions of online processing (see Appendix B).

Step 2 -Metaphor Identification
Identification of metaphorical expressions was carried out following Cameron and Maslen's (2010) modified version of the Pragglejaz Group's (2007) Metaphor Identification Procedure.The selected passages were reread, and possible metaphorical expressions were identified.These were checked for: (1) meaning in the discourse context; (2) the existence of another, more basic meaning; (3) an incongruity or contrast between these meanings and a transfer from the basic to the contextual meaning (Cameron & Maslen, 2010).
This can be illustrated using the first highlighted expression in Figure 2. The description of words which 'refuse to come' has 1) the meaning in context of attempted word-finding, 2) a more basic meaning of a person or other agent being Step 1 Step 2 Step 3

Figure 2. Illustration of Metaphor Identification and Coding into
Vehicle Groups 1 & 2 using an Extract from Wulf (1973) unwilling to physically move, therefore there is 3) a contrast in meaning between the mental processes of word-finding and working memory and the actions and attitudes of an agent.There is a transfer in meaning from the latter to the former through conceptualisation of the words as animate individuals, the mind as a place, and so on.All expressions which met these criteria were entered into a spreadsheet.
The identification of a more basic meaning inevitably involves a judgment by the researcher.The intuitions of the native English-speaking researchers, consultation of dictionaries and etymologies, and comparison of frequency of different usages in Internet search engine results, were used to increase the reliability of this judgment.This procedure was performed independently by at least two researchers for each text, with group discussion of unclear cases and documentation of the principles on which these cases were decided to ensure consistency.The two versions were then combined with further discussion of any points of disagreement.The first author checked each text and its data against the final document of agreed inclusion principles.

Step 3 -Coding of Vehicle Terms and Topics
The words used metaphorically (e.g., 'rounding up') are referred to as the 'vehicle' terms.These terms were coded into semantically related 'vehicle groups.'This was done at two levels of generality following Cameron and Maslen (2010): 'Vehicle Group 1' remaining as close to the specific meaning of the term as possible, and 'Vehicle Group 2' creating broader overarching semantic categories.If a metaphorical expression could be included in two categories, it was duplicated and coded in each.Figure 2 shows an extract of the coded data.This process was carried out iteratively and collaboratively by at least two researchers, including the first author, with consensus decision-making on points of disagreement.Following the coding of the data into Vehicle Groups, the data was also coded by topic, according to predetermined broad categories of language modality or level: spoken language, thinking/cognition/state of mind, reading, writing/typing, understanding spoken language, role/communicative ability, self-monitoring, non-verbal communication, and discourse/dialogue.

Step 4 -Description of Systematic Metaphors
A subset of the data was extracted for this analysis which addressed word production (this was not restricted by topic codes, as items coded as topics other than spoken language were relevant, for example, writing/typing and discourse/dialogue).The description of systematic metaphors was carried out by the first author through repeated examination of this narrower dataset, sorted by vehicle group, topic and author, and through reference to context, pattern of use and discourse function in the entire texts and in the entire dataset (for example to examine similarities with the descriptions of receptive language).Figure 3 shows multiple vehicle terms which were coded as the same Vehicle Group 2, of 'personification/animation.'As can be seen from these examples, there is not a straightforward translation of vehicle group codes into systematic metaphors, as different target domains (mental processes, body parts, linguistic units) are mapped using the same source domain (personification).

Step 5 -Interpretation of Systematic Metaphors
The interpretation of systematic metaphors in the light of psycholinguistic theory was carried out by the first author.These judgements are based on clinical and theoretical knowledge, and on the partial objective information provided in the accounts.In some instances the process was straightforward, where the authors themselves explicitly mentioned the symptom described, for example with more than one author linking descriptions of APHASIA AS SILENCE to "a lack of inner speech."In cases where the connection is more speculative, multiple examples are provided to support these interpretations.The application of the systematic approach of metaphor-led discourse analysis to a heterogenous range of accounts means that the results produced in this way can form the basis of hypotheses to be tested in future work.

Results and Discussion
The research aims to systematically analyze written accounts by people with aphasia to understand their subjective experiences of word production difficulty through metaphorled discourse analysis, to compare these subjective accounts to existing psycholinguistic models, and to identify symptoms not currently included in these models.To achieve this all descriptions of language processing or impairment were excerpted for analysis, which was carried out through the identification of metaphor vehicle terms.These vehicle terms were coded into vehicle groups, first at a fine-grained level of coding which remained as close as possible to the text, ('Vehicle Group 1'); and then at a more general level to allow for a broader understanding of the systematic use of related metaphors throughout and across the texts ('Vehicle Group 2').Both of these levels of coding were used to sort the data for repeated examination in the description of systematic metaphors.The coded Vehicle Group 2 categories are reported here to give a descriptive quantitative overview of the data before the qualitative exploration of systematic metaphors.Two distinct types of systematic metaphor emerged in the analysis, with related conventional systematic metaphors frequently recurring throughout all or most of the accounts, and some more

Quantitative Overview
Of 8148 total vehicle terms relating to language processing, 4080 metaphorical expressions addressed word production.There were 34 Vehicle Group 2 categories, with PHYSICAL OBJECTS/STRUCTURE predominating. Figure 4 shows the distribution of the 22 most frequent Vehicle Group 2 categories, including all of those discussed in this paper.Table 1 shows a breakdown by author of the most frequent twelve groups, which each contained over 100 instances across texts, and together accounted for 78% of expressions.There was consistency across the texts in the distribution of the most frequent vehicle groups, with the exception of Douglas (2002) who used a low absolute number of metaphorical expressions.

Qualitative Analysis -Systematic Metaphors
To explore how people with aphasia conceptualise and reason about their word production difficulties, eight systematic metaphors used to describe word production and its impairment are described.This was done on the basis of repeated examination of the data, sorted in different combinations by Vehicle Group 1, Vehicle Group 2 and author, with frequent reference to the context of the vehicle terms in the books.Five conventional metaphors were used across all or most accounts.The conventional metaphor of  VIEWPOINT/VISIBILITY "I have difficulty obtaining a general overview of my mental field" (Lubbock,p. 33) variations of this metaphor used to emphasise particular aspects of experience (e.g., WORDS AS LOST/STOLEN POSSESSIONS).Other conventional metaphors for cognition (Lakoff, 1994) used across all or most accounts were WORD PRODUCTION AS A JOURNEY/HUNT, APHASIA AS BODILY IMPAIRMENT, APHASIA AS FRAGMENTATION AND PERSONIFICATION OF SELF and SELF AS MACHINE/COMPUTER.These conventional metaphors were used to describe symptoms and to make fine-grained distinctions between different levels of representation and types of disrupted processing.Three novel metaphors used in an overlapping subset of the accounts to describe a pervasive qualitatively different experience arising from certain specific symptoms: APHASIA AS SILENCE, APHASIA AS A SPIRITUAL EXPERIENCE and APHASIA AS A DISMEMBERED TREE.

Conventional Systematic Metaphors
The majority of vehicle terms made use of conventional metaphors, which were used recurringly throughout all of the texts, with the 'conduit metaphor' COMMUNICATION AS PHYSICAL TRANSFER (Reddy, 1979) making up around half of all instances.To describe the nuances of word production difficulties these conventional metaphors were adapted in various ways: made more specific, elaborated in unusual ways, and idioms were made more vivid and explicitly metaphorical.Here the four most frequent metaphors which occurred in the book are described (including several different variations of the first of these).

WORD PRODUCTION AS MOVING AN OBJECT OUT OF A CONTAINER.
Unsurprisingly the conventional representation of COMMUNICATION AS PHYSICAL TRANSFER predominated, primarily as WORD PRODUCTION AS MOVING OBJECTS OUT OF A CONTAINER and WORD PRODUCTION AS MOVING OBJECTS OUT OF THE BODY.Details of the impairment could be mapped through the specification of particular properties of WORDS AS OBJECTS, and MIND AS CONTAINER.Numerous variations of this metaphor occurred, which are explored in detail below.This metaphor and its close variants are used by most authors to describe the initial symptoms or the acute experience of aphasia.Some authors, such as Schultz, retain this metaphor throughout their accounts.However other authors, like Broussard, explicitly reject it in favour of a different metaphor which better maps the nuances of their symptoms; yet others, like Taylor, make gradual increasing use of alternative metaphors throughout their description of recovery.

WORDS AS LOST/STOLEN POSSESSIONS and MIND AS EMPTY CONTAINER.
These related metaphors were primarily used to describe the early experience of word production difficulties in acute aphasia, and in the context of describing the severity and strangeness of the experience to others, drawing an explicit contrast with more ordinary experiences of language difficulty.While these are both used idiomatically in English to describe transient linguistic or cognitive difficulties, for example, "robbed of speech," "my mind is a sieve," several accounts used more emphatic descriptions of complete loss and emptiness, using words such as "void," "chasm" and "abyss."

MIND AS CLOSED/LOCKED CONTAINER.
An alternative way to describe the overall experience, or the severity of symptoms, through metaphors of objects and containers was through representing the container as closed or locked, that is as a difficulty with access.Taylor and Resch made ongoing use of this metaphor, describing changes in recovery using metaphors of specific containers, 'filing cabinets' and a 'black box' respectively (see Table 2).Through these metaphors Taylor describes both online word production, and the process of rehabilitation in terms of finding and tidying files, and Resch describes the changes in her word production abilities in terms of the properties of the container.

WORD PRODUCTION AS MOVING OBJECTS THROUGH THE BODY. This metaphor is
used by all of the authors, often to describe the sequential nature of word production, with a contrast made between "head" vs "mouth" to describe a subjective awareness of distinct stages of processing (see Table 3).There does not appear to be a precise and consistent set of mappings between level of impairment and place in body, but there is a distinction made between relatively more central lexical processes being described as in the 'head' or 'brain,' and more peripheral sub-lexical ones as in the 'mouth' or 'throat.'

WORDS AS MANIPULABLE, MESSY OR DISGUISED OBJECTS.
The ability to use a word was often described in terms of the manipulability of an object, particularly through 'grab' or 'grasp.'This metaphor was used more for receptive than productive language, but was sometimes used to describe difficulty in maintaining a representation in working memory, or to describe a subjective awareness of self-monitoring difficulties in language production (i.e., aspects of word production which may involve receptive language processes).Manipulability was also used to describe articulatory difficulties in terms of the ease of moving an object from the mouth.Another way in which more detail could be given about the nature of the impairment was through the addition of details about the properties of the physical objects.Metaphors of 'jumble' and 'clutter' were often used to describe difficulty with identifying and using a target 'idea' or 'fact' from various alternatives, suggesting selection difficulties at a lexical semantic level.This metaphor was, less frequently, applied to other levels as well, with West using it to describe difficulty selecting the required sublexical units.Other elaborations of WORDS AS OBJECTS which similarly describe difficulty distinguishing words from one another were the use by West and Broussard of WORDS AS DISGUISED and WORDS AS (WRONG) CONTAINERS to describe a mismatch of semantic and phonological representations.Examples of metaphors which attribute additional properties to WORDS AS OBJECTS are given in Table 3.

WORD PRODUCTION DIFFICULTIES AS A LOSS OF SOLIDITY.
With successful word production represented as the transfer of a physical object, there were mappings available for word production difficulties through metaphors of a loss of solidity.There were two groups of such metaphors: melting or disintegration, and evaporation (see Table 3).Melting or disintegration was used to describe the erroneous production of a successfully retrieved wordAn object that melts or disintegrates is something still perceptible to others, but which has changed shape and is difficult to manipulate, thus providing an apt mapping for the production of phonological paraphasias or articulatory difficulties.In contrast, when something evaporates, it is not perceptible.This metaphor was used to describe a difficulty prior to speech, in which no attempt at production was made.

WORD PRODUCTION AS A JOURNEY/HUNT. Metaphors of animals and people moving
through landscapes were used by most authors (see Table 4).This metaphor is similar in its basic logic to WORD PRODUCTION AS MOVING AN OBJECT OUT OF THE BODY, but with additional mappings available in the scaled-up metaphorical environment.This was used to describe the effortful and time-consuming nature of word retrieval in terms of distance or barriers in the landscape; the possibility of partial activation as being 'near' a word; internal selfcuing and circumlocution as finding alternative routes; and neuroplastic improvement as the repeated travel of a path creating a better path.The animation of words provides mappings for processes occurring outside of the person's sense of control, often in terms of hunting or herding disobedient or reluctant animals or people, a metaphor used by West to describe disordered activation and production of speech errors.A more idiosyncratic metaphor of animacy was Mills' descriptions of sleeping or drowsy 'thoughts' or 'ideas,' possibly reflecting a subjective difference arising from the different aetiology of TBI.
APHASIA AS BODILY IMPAIRMENT.Distinctions between more central and more peripheral difficulties could also be made using a metaphor of WORD PRODUCTION DIFFICULTIES AS BODILY IMPAIRMENT, or as a 'problem' with a particular body part (see Table 5).The use of these metaphors of the 'place' of the difficulty as a way to metaphorically map different levels of impairment is made especially clear by the authors who use this metaphor to describe difficulty at multiple stages of word processing.This metaphor also provided a mapping for the description of difficulties with self-monitoring stages of processing as a problem with 'hearing,' used by Schultz, Broussard and Marks.A different use of this metaphor was the description of MOTOR PLANNING/ ARTICULATORY DIFFICULTIES AS PROBLEMS WITH WALKING, with the self, the tongue, speech activity, or words themselves are represented as the person having difficulty walking.This metaphor could also describe successful speech as running or walking, compensatory strategies as aids to walking, and in one instance as a phenomenologically distinct form of self-monitoring of errors where in contrast to the more common descriptions of 'hearing' an error, Broussard (2016) uses the metaphor of an interrupted motor movement: "Not unlike stubbing a toe, I could feel it . . . in the same way I could feel an error" (p.83).

APHASIA AS FRAGMENTATION AND PERSONIFICATION OF SELF and SELF AS MACHINE/COMPUTER.
A different metaphor type with which difficulties were related to different body parts was the fragmentation and personification of parts of the person (see Table 6).'Brain,' 'mind,' 'mouth' were all personified, as were mental processes and aspects of language, such as 'inner voice,' and 'monitor.''Brain' or 'mind' were used in juxtaposition with the self and were described as carrying out high level cognitive tasks that would in conventional language be ascribed to the self, as in Schultz' "my head slowly and deliberately thinks out my condition."These uses of personification were often used to describe complicated disruptions to feelings of agency and control vs automatic processes occurring outside the self in speech production and self-monitoring.MIND AS MACHINE/COMPUTER was similarly used, though less often, to describe similar disruptions to a sense of agency.These metaphors were also used by Lubbock to describe the variability of his symptoms, as he describes different aspects of his language 'stalling' and 'glitching' or 'working automatically' at different times.

Novel Systematic Metaphors
While most of the metaphors used drew on conventional metaphors for communication, a subset of authors used novel metaphors to describe particular symptoms.A small range of such metaphors occurred, with overlap in the authors who used them.Their pattern and function was different than the recurring use of conventional metaphors, as they appeared less often, usually in extended passages, were often marked as important or metaphorical by the author, and in the context of describing the difficulty of conveying their experience, or the strangeness of their symptoms.

APHASIA AS SILENCE and APHASIA AS A SPIRITUAL EXPERIENCE.
Several authors used metaphors of silence to describe a lack of inner speech (see Table 7).Taylor and Marks describe this state of mind in the greatest detail and accord it the most importance.Resch uses the same metaphors, although she is describing an apparently milder language impairment and does not link these metaphors as clearly to specific language processes.West and Wulf make briefer use of them in reference to the acute stage of aphasia, as does Lubbock towards the end of his account of increasing loss of language.The authors who used this metaphor also described this experience of silence as a positive mental state, often using spiritual or religious terms, and sometimes involving a sense of expansion, unity or fluidity and bliss.The return of inner speech is described as, at least in some ways, a painful loss, and a return of 'noise' and anxiety.

APHASIA AS A DISMEMBERED TREE. The novel metaphor APHASIA AS A APHASIA AS
A DISMEMBERED TREE was used by Marks, Resch and Maria, and elaborated in detail by Maria and Marks, who used in extended descriptions, in the context of rejecting or elaborating on conventional metaphors suggested by others (see Table 7).In describing how the parts of the previously entire entity are still present, but the connections between them have been damaged or destroyed, this metaphor describes the subjective experience of difficulty in making semantic associations.Marks, Maria and Resch all describe an effect of aphasia on cognition, initial receptive problems, and semantic paraphasias.

Relating Metaphors to Psycholinguistic Models/Theory
The symptoms described using these metaphors can be interpreted in light of psycholinguistic theories and models.Table 8 summarises the hypothesised links between the systematic metaphors used and psycholinguistic descriptions.The results do not align strictly with one single theoretical approach or model, but indicate the importance of considering certain processes which may be better described by some models than others.Below is a summary of the main aspects of psycholinguistic theory which relate to the metaphors described above.

Modularity
The subjective symptoms align with a broadly modular language architecture even in breakdown.The level of detail varied, with a general distinction made between central and peripheral processes, which could potentially map onto more than one psycholinguistic level.However, even when multiple processing levels were affected, descriptions highlighted the distinct phenomenology of these impairments, as seen in West's account of acute global aphasia.These distinctions were made with reference to different places in the body (see Table 8, CONTAINERS AND OBJECTS and BODILY IMPAIRMENT).

Connectionist Processes
Despite these broad distinctions between modular levels, there were aspects of the subjective descriptions which reflect processing within, and interactively between, these levels.These can be seen through descriptions of difficulties with the activation, selection, and retention of phonological word forms as solidification or disintegration (see Table 8 CHANGE OF STATE, FLUIDITY, SOLIDITY).Partial activation, internal self-cuing, and neuroplasticity were also described, in particular using metaphors of travel through landscapes (see Table 8 JOURNEYING THROUGH A LANDSCAPE).

Cognitive Processes
Multiple authors highlighted difficulties with attention and cognitive processes as contributing to their word production difficulties; conversely the authors who described semantic and inner speech impairments described an impact of language impairment on cognition.Authors also described a disrupted sense of agency and conscious control over language production (see

Self-monitoring
Self-monitoring was primarily described through personification of aspects of the self (see

Inner Speech
The authors who described impaired inner speech identified this as the key feature of their aphasia, affecting cognition and sense of self, as well as affecting language production, engagement in rehabilitation and conceptualisation of 'recovery' (see

Final conclusions
We set out to 1) explore which metaphors are used to describe the subjective experiences of word finding difficulties in written accounts by people with aphasia, 2) to consider whether these descriptions reflect the details of the cognitive and linguistic impairment as described by psycholinguistic models, and 3) to consider whether particular processes or symptoms which are described are not accounted for in some models or theories.These research questions are addressed in order below.Predominantly conventional metaphors for communication and cognition are used to describe the subjective experience of word finding difficulties in aphasia.That is, the everyday metaphors that we use can be extended and elaborated to map many aspects of disrupted word production in aphasia.Certain symptoms were described using more novel metaphors, which were also explicitly highlighted as being unusual, pervasive and hard to explain to others.These novel metaphors were used by an overlapping subset of authors who experienced impairment of inner speech, cognitive involvement, difficulty making semantic associations, and initial receptive symptoms.It is a clinically important finding that some people with aphasia experience a profound alteration of experience, which is hard to describe even after recovery of adequate language (requiring creative use of novel metaphors), and which affects motivation, sense of self and attitudes about recovery and language use.A reported lack of inner speech was particularly noteworthy as a symptom which multiple novel metaphors were employed to describe, in strikingly consistent ways across several accounts.The authors who emphasized this symptom described not merely recovering or losing language, but renegotiating their relationship with language and verbal thought, and confronting spiritual and existential questions, regardless of their premorbid beliefs or interests.It is important for speech and language therapists to understand that some people with aphasia may be preoccupied with such experiences and questions.It may lead to complex emotions about language recovery and engagement in therapy as described in these accounts.It is also useful for psychologists, hospital chaplains, and the family and friends of those with aphasia to be aware of as an important experience which may affect a person's values and personality.Mumby and Roddam (2021) provide a valuable tool to support people with aphasia in communicating about these topics.
The details of the subjective experience of word finding difficulties aligns with psycholinguistic and neuropsychological theory.Even when multiple levels of processing were affected for an individual, the subjective description distinguished these different aspects of the impairment.The level of detail was not always as fine-grained as psycholinguistic theory: often a general distinction was made between more central and more peripheral processes, with these experiences possibly mapping onto more than one psycholinguistic level.However, at times distinctions were made which map onto specific levels of processing and even the way in which processing is disrupted within that level.The differences between difficulties with activation, selection and retention of a phonological word form were consistently described by multiple authors.It is useful for clinicians to know that impairment of different stages of word production difficulties in aphasia can be experientially very distinct, even when behaviourally similar and when there are multiple severely impaired stages.Careful attention to the way that people with aphasia describe their symptoms can provide a valuable source of information to be integrated with objective measures.
This suggests that clinical practice should make wider use of more complex and synthesising models of word production such as Roelof's WEAVER++/ARC (2014), which includes processes of attention and self-monitoring, goal-referenced control, selection, and spreading activation.Assessment with reference to a model which allows for description of impairment to these processes has the potential to improve understanding and communication about the aspects of impairment which are salient to people experiencing aphasia.Such a model also draws our attention to the role of consciousness in word production, as a part of typical processing, as potentially disrupted by impairment, and as being able to contribute to compensation or recovery.

Limitations and Future Directions
Certain limitations arise unavoidably from the use of published autobiographies.The primary limitation is that these accounts were not produced in controlled research settings, and the degree of editorial support or alteration is unknown.The sample of people with aphasia included in the analysis is not representative, as the writing and publishing of a book requires considerable internal and external resources (language and cognitive skills, time and energy, access to publishers or the knowledge or support to publish independently, etc.).Triangulation with other methods will strengthen these results, such as interview-based metaphor analysis, which can use rigorous controls, obtain objective behavioural and neurological data, and would allow for checking of codes with participants.
Certain aspects of language processing are more available to conscious awareness than others.For instance there were numerous descriptions of self-monitoring consistent with a receptive-language based theory of self-monitoring, but this may be because a methodology which takes the reported contents of consciousness as its data finds a fit in models which incorporate consciously accessible representations.The single instance of a description of self-monitoring which is consistent with the alternative account of a production-based monitor was provided by Broussard, who describes a particularly deliberate and nuanced process of reasoning about the nature of his language processing, and so might be expected to report some aspects of word production which are less apparent to introspection or less amenable to description.Thus on the basis of this data we may argue for the inclusion of self-monitoring through receptive language networks in clinical models as a salient aspect of language production, but not against the possibility that much self-monitoring may be production-based.
Further analysis of this data will explore in greater depth the role of inner speech, and the use of metaphors for metacognition about impairment and recovery.Future work will explore the use of metaphor to describe symptoms of word production difficulties in aphasia through other methodologies, such as semi-structured interview, group discussions, questionnaire or through visual materials.Such work would also allow for neurological information to be obtained and incorporated into the analysis.
Continued research in this area would support the development of a communication tool for the visual and verbal presentation of metaphors to improve clinical communication between speech and language therapists and their clients with aphasia.The 'Metaphor Menu for people living with Cancer' (Semino, 2019) provides a model for such a tool.This research illustrates that careful attention to the way that people with aphasia describe their symptoms can provide a valuable source of information to be integrated with objective measures, in research and in clinical practice.(Continued) Appendix B Overview of how methodological trustworthiness was ensured at each stage of the analysis.
Identification of metaphors was carried out following the procedure outlined by Pragglejaz Group (2007) and Cameron and Maslen (2010), coding of Vehicle Groups and topics and description of systematic metaphors following the recommendations of Cameron, Maslen & Low (2010), and recommendations on working on large amounts of metaphor data were incorporated following Maslen (2010).
Step 1: Familiarisation with Texts and Selection of Descriptions of Language Processing Familiarity with the texts through reading and rereading was a necessary foundational step in reducing researcher expectation and bias.It was crucial initially for understanding idiosyncratic expressions used to communicate about aspects of language processing.These expressions can involve unique metaphors, abstract concepts, or shorthand phrases that encapsulate complex, contextually rich aspects of their experience of aphasia.Thus to identify relevant sections for analysis it was necessary to understand that for example, any reference by Marks (2017) to 'the Quiet,' by Wulf (1973) to 'the extra room in my head,' or by West (2008) to 'the BBC man' refer to aspects of language processing.The selection of relevant passages before the identification of potential metaphors reduced the risk of missing metaphors due to researcher expectations (Cameron & Maslen, 2010).An inclusive approach was taken to selection of relevant passages as it was not always possible to draw a clear distinction between descriptions of impact and of the impairment.The inclusion of disputed cases allowed for as comprehensive a dataset as possible to be compiled, while reducing the risk of missing relevant metaphors and lessening the influence of researchers' subjective decision-making on case-by-case decisions.
Step 2: Metaphor Identification Trustworthiness was maximised through initial training, consensus decision-making and documentation of decisions, and a final check for consistency, following the recommendations of Cameron and Maslen (2010).Two group workshops with the research assistants were conducted, followed by individual training sessions and checking of expressions identified until a reliability rate of >80% for twenty consecutive items was reached (with this process repeated for each book).Once a consistent standard of identification was established, regular group discussions were implemented to maintain this consistency.Decisions on inclusion and exclusion criteria which resulted from problematic cases were recorded (see Appendix C), and the first author carried out a final check using these criteria across all of the sources.Familiarity with the texts also supported the next stage in the analysis as some conventional metaphors may not be immediately conspicuous, or their metaphoricity may not be apparent from a single instance.However, when such metaphors appear repeatedly, or are in some instances elaborated and more obviously metaphorical, this provides justification for their inclusion.
Step Reliability in coding of vehicle groups was maximised through collaborative decision-making in regular group discussions between the first author and research assistants, and cross-checking of all data by at least two researchers, with each book checked by the first author and at least one research assistant (Cameron, Low & Maslen, 2010).A two-level coding system was used.'Vehicle Group 1' codes used language close to that of the source texts, whereas 'Vehicle Group 2' provided a higher level of abstraction.This structure helped ensure that broad categories or overinterpretations did not overshadow the nuance of the original data.It also allowed easy crosschecking of 'Vehicle Group 2' labels against 'Vehicle Group 1' codes, ensuring they were true to the original expressions.We grouped metaphor vehicles before adding topic codes, following Cameron Low and Maslen's (2010) recommendation that this helps researchers to "avoid trying to second guess what the speakers meant and concentrate on the words they actually said." Step 4: Description of Systematic Metaphors The first author led the description of systematic metaphors, due to the need for familiarity with the entire dataset and the full autobiographical accounts.To triangulate this process, another researcher conducted an analysis on four accounts, identifying five systematic metaphors which were used to describe word production.This independent analysis agreed on the two most prevalent metaphors (with minor variation in wording and emphasis: LANGUAGE/WORDS AS OBJECTS, and WORDS AS SPATIALLY LOCATED).SELF AS MACHINE/COMPUTER was also found, as in the current study; the remaining two metaphors WORD-FINDING AS FIGHT, and SELF AS CHILD, are present in the larger dataset but are not discussed in the current study as they did not feature as prominently as the metaphors which are included.Thus there was considerable agreement across these two independent interpretations of the data, with agreement on the two most prevalent metaphors, and differences in the description of less common metaphors which are consistent with use of a narrower dataset.Following Maslen's (2010) recommendation that "it is important to avoid treating the dataset as contiguous data, losing touch with the contextual meaning of its original context," the original full accounts were regularly consulted during this stage of analysis.Care was taken not to introduce theoretical preconceptions into the analysis, with the interpretation of systematic metaphors carried out without reference to specific models of word production.The description of systematic metaphors were carried out without the imposition of preconceived theoretical models.
Step 5: Interpretation The first author engaged in regular discussions with the second and third authors to reflect on her interpretation of the data, and to address issues related to expectation and bias.To minimise the influence of theoretical preconceptions, the description and interpretation of systematic metaphors was carried out without prior commitment to one or more specific models of word production.Rather, psycholinguistic concepts were applied in a piecemeal and pragmatic way in response to the data.That the findings provide support for symptoms represented by different types of model provides evidence that this was effective.The emergence of themes which were not anticipated, and which were not pre-existing research interests of the authors (such as spiritual experiences relating to inner speech) also reflect success in reducing the impact of theoretical preconceptions on description of systematic metaphors.
Many of the autobiographical accounts include objective reports or descriptions of their symptoms.While these do not constitute directly collected objective data, they do provide some evidence for the validity of the interpretations given here.A further source of internal evidence is in the robustness of the relationship between the vehicle groups identified, and the specific aspects of word production which they describe.Cameron, Low and Maslen (2010) state that "the more robust the relationship [between semantically similar metaphor vehicles and the topics they express], the stronger the claim that can be made about the underlying factors it reveals."That there is high consistency within and across accounts between use of particular vehicle groups and the aspects of word production being referred to is evidence of the systematicity of the metaphors described machine/computer: car/train/vehicle, computer/info processing/circuitry, mechanical/electrical/ hydraulic violence/competition: battle/war; blast/barrage, dangerous, defeat, defence/siege, fatal/ lethal, force/seize, game/sport/physical effort, physical fight/assault/struggle, retreat/ advance/battlefield, revenge, shock/stun/silence, soldier/army, survive, victim, weapons/ resources clarity/darkness: clarity/lucidity/light, darkness, fog/blur viewpoint/visibility: fractured/distorted vision, insight/see in, invisibility, obscured vision, perspective/focus, reflect/mirror

Figure 3 .
Figure 3. Examples of Data Sorted (by Vehicle Group 2, Author, and order of occurrence) for description of Systematic Metaphors

"
"a "third person" talking to a "first person""(Broussard, p. 57)    "that word . . .would slide into a more accessible area for future use" (Maria) . . .belong to another world"(Mills, p. 361) LANGUAGE; CREATIVE WRITING "I am learning English as a second language" (Schultz, p. 157) "words were found in funny little places called 'the back way'"(Wulf, p. 54)    "my mind and speech had been short-circuited" (Maloney, p. 46) "my outright war with the other consonants and vowels"(West, p. 165) Cannot find the words through the fog"(Schultz, p. 8

Table 1 .
Twelve most frequent Vehicle Group 2 Codes, that is, the broader semantically related groupings of vehicle terms which allow for comparison across accounts.
Figure 4. Total count of the twenty-two most frequent Vehicle Group 2 Codes "I've been hunting the word capillary"(Wulf, p. 86)

Table 2 .
Variations of the conventional metaphor WORD PRODUCTION AS MOVING AN OBJECT OUT OF A CONTAINER used across all or most accounts to describe the initial or overall experience of word production difficulties, or to emphasize their severity or strangeness.Even though my brain remained lined with filing cabinets it was as if all the drawers had been slammed shut.the word to find those files, open them and remember.Taylor 85 A black box, words locked inside, no door, only a few words drifting outside Resch 7 But most of the pieces are still in the black box.I clutch at the door, wrestle to pry it open . . .reach in, and scramble around to find pieces to fit.
30Even this far out from the stroke, I can't open the black box of words at will.

Table 3 .
Variations on the conventional metaphor WORD PRODUCTION AS MOVING AN OBJECT OUT OF A CONTAINER used to describe more specific aspects of word production difficulties.

OBJECTS MOVING THROUGH THE MIND AND BODY
the spelling of his name finds its place in my head then a place in my mouth Schultz 96 the laborious task of forcing an idea through my caved-in brain to the paper Wulf 122 until the word finally rattled off my tongue correctly: Tarantula!Marks 82 I ...set a cupped hand in front of my face as though I were waiting for a word to fall into it Maria 85 you disentangle the least bit of wiry fluff that has been haunting your tongue for half an hour , in my mind, the words I'm using ready for the next time they slip away Lubbock 111 However, I couldn't grab it in time to say it in a normal conversation.Broussard 82 I cannot grasp [numbers] reliably when I am saying them or ...hearing them.Schultz 133 you disentangle the least bit of wiry fluff that has been haunting your tongue for half an hour disguised as a word is not a word at all.For instance Robitracing or my own word Turps.West 51 I heard a cry of Inderal pretending to be the anti-hypertensive Cozaar and of Cozaar pretending to be Inderal.(Or even of Cozaar pretending to be Cortazar!) Mills 105words will cascade ...from my mind, and if not captured in type instanter, it may be too late.They were such right words -why must they evaporate so soon?Wulf 72I tried to say things (and sometimes came close) that had a "fading" quality Broussard 66

Table 4 .
Examples of the conventional metaphor WORD PRODUCTION AS A JOURNEY/HUNT, with an emphasis on MIND AS LANDSCAPE used to describe effort, difficulty and use of strategies, and an emphasis on WORDS AS ANIMATE to describe a lack of control, or production of errors.First stumbling all about the phrase . ..then stumbling further as I worked each word in turn through the mazes of my corrupt language.continueto use my voice,find pathways and plod over them often enough to make them visible to thoughts sent from my mind and less likely to be erased from lack of trodding Wulf 55When I couldn't find a word, it felt like a bridge leading to the word had been burned . . .I could get there (eventually) by looking for other (less complicated) bridges.
Frustration ...with fatigue ...chokes out my flimsy path to words like brambles concealing roses.Resch 36WORD PRODUCTION AS A JOURNEY/HUNT (WORDS AS ANIMATE)Every time I settled on an idea that would give me courage, it flew from me at top speed and left me floundering in its wake like a rodeo dancerWest 119Of all those [words] milling around in self-evident joy, it's hard from my point of view to distinguish them from each other.West 126decided to go on a slightly different route from the one prescribed for it West 155 the misbehavior of ordinary words ...some of the right words cannot be said and others take their places without mercyWest 175to persuade the right words, or any words, to leave their cozy nest in my mind to traverse the rocky road through my brain to the outside world Wulf 134The ideas are there, but they're . . .well, asleep, dormant, waiting for a transfusion Mills 297

Table 5 .
Examples of conventional metaphor APHASIA AS BODILY IMPAIRMENT used to specify and differentiate impairment at different stages of word production.

Table 6 .
Examples of APHASIA AS FRAGMENTATION AND PERSONIFICATION OF SELF and SELF AS MACHINE/COMPUTER

AND PERSONIFICATION OF SELF (BRAIN/MIND/MENTAL PROCESS PERSONIFIED)
While my monitor is usually on duty there are too many times when fatigue whispers "don't pay attention to him" and I don't Talking is tiring and if done too much at a time muscles controlling speech get very uppity and scream at me "We're not gonna struggle thru one more word" back through re/who're/your head as a check spell?Schultz 173 then I put my brain back on scan and eventually I access the right data Taylor 48 thoughts must be programmed through one's brain Wulf 125 The manual process slowly disappeared as the automatic sequence took over.Broussard 44 In order for me to say a word, my brain requires me to use a signal or a switch . . . to move my hand slowly across a table or across my other hand as I verbalize the word.
Maloney 51language seems to enact a signal or a mesmeric function, to the distress of the signaler or mesmeristWest 111

Table 7 .
Examples of novel metaphors APHASIA AS SILENCE, APHASIA AS A SPIRITUAL EXPERIENCESPIRITUAL EXPERIENCE and the Quiet . . .was much more interesting than my medical state Marks 4 the Quiet had become harder to access as my inner and outer voices had become louder Marks 138 those little voices, that brain chatter . . .were delightfully silent.tree of wisdom which has been uprooted, dismembered bit by bit, branches, leaves, roots and fruit, and that the trunk has been totally severed.Like a trunk, branches, leaves, roots and fruits, words and phrases float around in my head.Like the tree, I am stripped of the branches of my speech.Resch 79 Aphasia was like a mad gardener that sliced the branches and limbs away from the trunk.This sparse topiary cut me off from my usual points of reference, keeping me from associating my thoughts with one another Marks 101

Table 8 .
Summary of the suggestions made in this study about how metaphors used in subjective descriptions may align with our objective understanding of word-production difficulties.

OR HERDING ANIMALS; JOURNEYING THROUGH A LANDSCAPE ONLINE WORD-PRODUCTION; WORD-PRODUCTION STRATEGIES
Table 8 COMPUTER OR MACHINE; FRAGMENTATION AND PERSONIFICATION OF PARTS OF SELF).

Table 8
FRAGMENTATION AND PERSONIFICATION OF PARTS OF SELF) and through metaphors relating to 'hearing' or 'catching' the words (see Table8CONTAINERS AND OBJECTS or BODILY IMPAIRMENT).A single instance of a different metaphor was used to describe self-monitoring, in Broussard's description of 'feeling' an error as being "Not unlike stubbing a toe" (p.83), which he contrasts explicitly with 'hearing' mistakes (see Table8BODILY IMPAIRMENT).