Pairs Of Items Research Articles

Enhanced Multi-Relationships Integration Graph Convolutional Network for Inferring Substitutable and Complementary Items

Understanding the relationships between items can improve the accuracy and interpretability of recommender systems. Among these relationships, the substitute and complement relationships attract the most attention in e-commerce platforms. The substitutable items are interchangeable and might be compared with each other before purchasing, while the complementary items are used in conjunction and are usually bought together with the query item. In this paper, we focus on two issues of inferring the substitutable and complementary items: 1) how to model their mutual influence to improve the performance of downstream tasks, 2) how to further discriminate them by considering the strength of relationship for different item pairs. We propose a novel multi-task learning framework named Enhanced Multi-Relationships Integration Graph Convolutional Network (EMRIGCN). We regard the relationship inference task as a link prediction task in heterogeneous graph with different types of edges between nodes (items). To model the mutual influence between substitute and complement, EMRIGCN adopts a two-level integration module, i.e., feature and structure integration, based on experts sharing mechanism during message passing. To obtain the strength of relationship for item pairs, we build an auxiliary loss function to further increase or decrease the distances between embeddings of items with weak or strong relation in latent space. Extensive experiments on both public and industrial datasets prove that EMRIGCN significantly outperforms the state-of-the-art solutions. We also conducted A/B tests on real world recommender systems of Meituan Maicai, an online supermarket platform in China, and obtained 15.3% improvement on VBR and 15.34% improvement on RPM.

Evaluation of the Recommendation Effect of Individual Interventions in a Recommender System

Collaborative filtering is a recommendation model that evaluates and recommends items that match the preferences of each user and with high purchase probability. However, some items are regularly purchased and do not necessarily require recommendation. Thus, identifying items with a high recommendation effect is one of the challenges when using recommendation models. The recommendation effect can be regarded as an intervention effect in causal inference by considering the recommendation of an item as intervention. The intervention effect of individual users can be estimated using models such as counterfactual regression (CFR). Nonetheless, because this model can only manage one type of intervention, it cannot be easily and directly applied to recommender systems when many recommended items appear as intervention. In this study, we extend this model to estimate the effects of individual interventions in multiple interventions using a single CFR by combining user and item features to form covariates of user–item pairs, thereby allowing the recommendation effects to be estimated for all user–item pairs. We demonstrate the effectiveness of the proposed method through experiments using artificial data.

Sequential involvements of the perirhinal cortex and hippocampus in the recall of item-location associative memory in macaques.

The standard consolidation theory suggests that the hippocampus (HPC) is critically involved in acquiring new memory, while storage and recall gradually become independent of it. Converging studies have shown separate involvements of the perirhinal cortex (PRC) and parahippocampal cortex (PHC) in item and spatial processes, whereas HPC relates the item to a spatial context. These 2 strands of literature raise the following question; which brain region is involved in the recall process of item-location associative memory? To solve this question, this study applied an item-location associative (ILA) paradigm in a single-unit study of nonhuman primates. We trained 2 macaques to associate 4 visual item pairs with 4 locations on a background map in an allocentric manner before the recording sessions. In each trial, 1 visual item and the map image at a tilt (-90° to 90°) were sequentially presented as the item-cue and the context-cue, respectively. The macaques chose the item-cue location relative to the context-cue by positioning their gaze. Neurons in the PRC, PHC, and HPC, but not area TE, exhibited item-cue responses which signaled retrieval of item-location associative memory. This retrieval signal first appeared in the PRC, followed by the HPC and PHC. We examined whether neural representations of the retrieved locations were related to the external space that the macaques viewed. A positive representation similarity was found in the HPC and PHC, but not in the PRC, thus suggesting a contribution of the HPC to relate the retrieved location from the PRC with a first-person perspective of the subjects and provide the self-referenced retrieved location to the PHC. These results imply distinct but complementary contributions of the PRC and HPC to recall of item-location associative memory that can be used across multiple spatial contexts.

An Iterative Three-Stage Neighborhood Search for Solving Precedence Constrained Agricultural Land Investment Problem

The use of neighborhood search techniques to address a practical issue faced by agricultural investors is examined in this study. The problem is named as agricultural land investment problem with precedence constraints and it has an essential impact on agriculture issues. The tackled problem can be viewed as a variant of the well-known classical 0-1 knapsack problem where precedence constraints are imposed on pairs of items. Precedence constraints take into account a precedence relation between items. This paper first simulates the considered problem as precedence constraints knapsack problem and presents a mathematical representation model. Then, an iterative three-stage neighborhood search method is proposed for optimizing the problem. The proposed method consists of three stages. First stage applies a greedy procedure in order to construct a feasible solution. Second stage applies local search procedures in order to enhance the quality of the solutions at hand. Third and last, in order to broaden the search space, a random neighborhood destruction approach is introduced. Finally, the effectiveness of the suggested approach is assessed and contrasted with the outcomes obtained by greedy and local search techniques. The presented method is competitive and efficient since it produces excellent solutions in a reasonable amount of time.

Agency as a bridge to form associative memories.

The perception of agency can influence memory when individuals feel their decisions exert control over their environment. While perceived agency has been shown to increase memory for items, most real-life situations are much more complex. Here, we examined how an individual's agency to influence the outcome of a situation affects their ability to learn associations between items that occur prior to and after a decision is made. In our paradigm, participants were told they were playing a game show where they had to help a trial unique cue, a "contestant," choose between three doors. On "agency" trials, participants were allowed to pick any door they wanted. On "forced-choice" trials, participants were instructed to select a door that was highlighted. They then saw the outcome, a "prize" that was behind the selected door. Across multiple studies, we find evidence for agency-related memory enhancements for contestants, a pattern that extended to contestant-prize, contestant-door, and door-prize associations. Additionally, we found that agency benefits for inferred cue-outcome relationships (i.e., door-prize) were limited to situations when choices were motivated by an explicit goal. Finally, we found that agency acts indirectly to influence cue-outcome binding by enhancing processes akin to inferential reasoning which associate information across item pairs containing overlapping information. Together, these data suggest having agency over a situation leads to enhanced memory for all items in that situation. This enhanced binding for items may be occurring by the formation of causal links when an individual has agency over their learning environment. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Psychometric properties of teacher classroom assessment literacy instrument using Rasch model analysis

<span lang="EN-US">A teacher classroom assessment literacy (TeCAL) instrument was developed to measure the level of teacher classroom assessment literacy in schools. TeCAL contains 66 multiple choice items with four options based on four constructs namely purpose, measurement, evaluation and use. Thus, this study aims to identify the psychometric properties of TeCAL using Rasch measurement model (RMM) analysis through Winstep software version 3.72.3. The findings show that the compatibility values of mean square (MNSQ) infit and outfit items ranged from 0.64 to 1.46 and 0.40 to 2.23, respectively. The value of MNSQ outfit was outside the set range, but still met the other fit statistics indicator which has a positive point measure correlation (PTMEA) value. In addition, the findings show that the empirical raw variance explained by measures is 38.2%. It was very close to the modeled value of 38.4% with the empirical unexplained variance in 1st contrast being 7.5% less than the maximum controlled 15%. Largest standardized residual correlations identified 10 pairs of dependent items to be less than 0.7. The person and item reliability index values were 0.94 and 0.89 with separation index values of 2.90 and 3.80, respectively. Overall, this psychometric analysis is crucial to ensure that the TeCAL instrument has good quality and meaningful to use.</span>

Modeling multi-attribute and implicit relationship factors with self-supervised learning for recommender systems

Interactions of users and items can be naturally modeled as a user-item bipartite graph in recommender systems, and emerging research is devoted to exploring user-item graphs for collaborative filtering methods. In reality, user-item interaction usually stems from more complex underlying factors, such as the users’ specific preferences. A user-item bipartite graph could be used to understand the differences in motivation. However, existing research has not clearly proposed and modeled the factors that affect the differences, ignoring the similarities between user pairs and item pairs, preventing them from capturing fine-grained user preferences more effectively. In addition to the two points mentioned above, most GNN-based models for recommendation have the following two limitations: First, the model’s accuracy depends on the number of observed interactions in the dataset. Secondly, node representations are vulnerable to noisy interactions. This work has developed a novel recommendation model called “Multi-Attribute and Implicit Relationship Factors With Self-Supervised Learning for Collaborative filtering” (MIS-CF), which explicitly proposes and models multi-attribute and implicit relationship factors for collaborative filtering recommendation. Meanwhile, an auxiliary self-supervised learning task is designed to help the downstream task optimize the node representation. MIS-CF aggregates multi-attribute spaces through the user-item bipartite graph and establishes user-user and item-item graphs to model the similar relationship information of neighbor pairs through a memory model. The self-supervised learning task generates contrastive learning via self-discrimination, thus mining the rich auxiliary signals within the data, improving the accuracy and robustness of our model. Moreover, the sparse regularizer is utilized to alleviate the overfitting problem. Extensive experimental results on three public datasets not only show the significant performance and robustness gain of the proposed model but also prove the effectiveness and interpretability of fine-grained implicit factors modeling.

Can EAT-10 Become EAT-5? Improving Measurement Efficiency of Dysphagia with Item Response Theory.

To assess: (1) the Eating Assessment Tool (EAT-10) with item response theory (IRT) to determine which individual items provide the most information, (2) the extent to which dysphagia is measured with subsets of items while maintaining precise score estimates, and (3) if 5-item scales have the differing discriminatory ability, as compared to the parent 10-item instrument. Prospectively collected data from 2,339 patients who completed the EAT-10 questionnaire during evaluation at a tertiary care otolaryngology clinic were utilized. IRT analyses provided discrimination and location parameters associated with individual questions. Residual item correlations were also assessed for redundant information. Based on these results, three 5-item subsets were further evaluated using item information function curves. Areas under receiver-operator characteristic curves (ROC-AUC) were also calculated to evaluate the discriminatory ability for dysphagia-related clinical diagnoses. Item discrimination parameter estimates ranged from 1.71 to 5.46, with higher values indicating more information. Residual item correlations were determined within item pairs, and location parameters were calculated. Based on these data, in combination with clinical utility, three 5-item subsets were proposed and assessed. ROC-AUC analyses demonstrated no significant difference between the EAT-5-Alpha subset and the original 10-item instrument for discriminating dysphagia as a primary diagnosis (0.88, 0.88). The EAT-5-Clinical subset outperformed the original 10 instruments in ROC-AUC for aspiration. The EAT-5-Range subset was significantly associated with problems with thin liquids. IRT analyses distinguished three proposed 5-item subsets of the EAT-10 instrument, supporting shorter survey options, while still reflecting the impact of dysphagia without significant loss of discrimination. 3 (Diagnostic testing with consistently applied reference standards, partial blinding). Laryngoscope, 133:3327-3333, 2023.

노년기 우울 증상 평가에서 주관적 척도와 객관적 척도 간 상관성 연구

Objective: The aim of this study was to investigate the correlation between subjective and objective depression rating scale for the elderly with depressive symptoms.
 Methods: Thirty three outpatients aged more than 55 with depressive symptoms were assessed with the Patient Health Questionnare-9 (PHQ-9) and the Montgomery-Asberg Depression Rating Scale (MADRS). Similar items from each scale were paired to investigate their correlations. The level of agreement between the PHQ-9 and the MADRS was also assessed for subjects by using discrepancy-score.
 Results: Correlation between total scores of the PHQ-9 and the MADRS was significant (r=0.85, p<0.001). There were significant correlations between the paired items except for the item about volition. These results were the same in the group whose total education years are 7 or more. However, in the group whose total education years are 6 or less, there was no significant correlation between the paired items about sleep, concentration and volition.
 Conclusion: These results demonstrate that using the PHQ-9 in elderly patients is reliable and education level might be considered when interpreting the scale.

Item response theory assumptions were adequately met by the Oxford hip and knee scores

ObjectivesTo develop item response theory (IRT) models for the Oxford hip and knee scores which convert patient responses into continuous scores with quantifiable precision and provide these as web applications for efficient score conversion. Study Design and SettingData from the National Health Service patient-reported outcome measures program were used to test the assumptions of IRT (unidimensionality, monotonicity, local independence, and measurement invariance) before fitting models to preoperative response patterns obtained from patients undergoing primary elective hip or knee arthroplasty. The hip and knee datasets contained 321,147 and 355,249 patients, respectively. ResultsScree plots, Kaiser criterion analyses, and confirmatory factor analyses confirmed unidimensionality and Mokken analysis confirmed monotonicity of both scales. In each scale, all item pairs shared a residual correlation of ≤ 0.20. At the test level, both scales showed measurement invariance by age and gender. Both scales provide precise measurement in preoperative settings but demonstrate poorer precision and ceiling effects in postoperative settings. ConclusionWe provide IRT parameters and web applications that can convert Oxford Hip Score or Oxford Knee Score response sets into continuous measurements and quantify individual measurement error. These can be used in sensitivity analyses or to administer truncated and individualized computerized adaptive tests.

Debiased Recommendation with User Feature Balancing

Debiased recommendation has recently attracted increasing attention from both industry and academic communities. Traditional models mostly rely on the inverse propensity score (IPS), which can be hard to estimate and may suffer from the high variance issue. To alleviate these problems, in this article, we propose a novel debiased recommendation framework based on user feature balancing. The general idea is to introduce a projection function to adjust user feature distributions, such that the ideal unbiased learning objective can be upper bounded by a solvable objective purely based on the offline dataset. In the upper bound, the projected user distributions are expected to be equal given different items. From the causal inference perspective, this requirement aims to remove the causal relation from the user to the item, which enables us to achieve unbiased recommendation, bypassing the computation of IPS. To efficiently balance the user distributions upon each item pair, we propose three strategies, including clipping, sampling, and adversarial learning to improve the training process. For more robust optimization, we deploy an explicit model to capture the potential latent confounders in recommendation systems. To the best of our knowledge, this article is the first work on debiased recommendation based on confounder balancing. In the experiments, we compare our framework with many state-of-the-art methods based on synthetic, semi-synthetic, and real-world datasets. Extensive experiments demonstrate that our model is effective in promoting the recommendation performance.

A New Questionnaire to Assess Respiratory Symptoms (The Respiratory Symptom Experience Scale): Quantitative Psychometric Assessment and Validation Study.

Smokers often experience respiratory symptoms (eg, morning cough), and those who stop smoking, including those who do so by switching completely to electronic nicotine delivery systems (ENDS), may experience reductions in symptoms. Existing respiratory symptom questionnaires may not be suitable for studying these changes, as they are intended for patient populations, such as those with chronic obstructive pulmonary disease (COPD). This study aimed to develop a respiratory symptom questionnaire appropriate for current smokers and for assessing changes when smokers stop smoking. The Respiratory Symptom Experience Scale (RSES)was derived from existing instruments and subject matter expert input and refined through cognitive debriefing interviews (n=49). Next, for purposes of the quantitative psychometric evaluation, the RSES was administered to smokers (n=202), former smokers (no tobacco use in >6 months; n=200), and switchers (n=208, smokers who switched to ENDS for >6 months), all of whom had smoked for at least 10 years (mean age 33 years). Participants, who averaged 62 (SD 12) years of age, included 28% (173/610) with respiratory allergy symptoms and 17% (104/610) with COPD. Test-retest reliability was assessed by repeat assessment after 1 week in 128 participants. A generalized partial credit model confirmed that the response options were ordered, and a parallel analysis using principal components confirmed that the scale was unidimensional. With allowance for 2 sets of correlated errors between pairs of items, a 1-factor graded response model fit the data. Discrimination parameters were approximately 1 or greater for all items. Scale reliability was 0.80 or higher across a broad range of severity (standardized scores -0.40 to 3.00). Test-retest reliability (absolute intraclass correlation) was good, at 0.89. RSES convergent validity was supported by substantial differences (Cohen d=0.74) between those with and without a diagnosis of respiratory disease (averaging 0.57 points, indicating that differences of this size or smaller represent meaningful differences). RSES scores also strongly differentiated those with and without COPD (d=1.52). Smokers' RSES scores were significantly higher than former smokers' scores (P<.001). Switchers' RSES scores were significantly lower than smokers' scores (P<.001) and no different from former smokers' scores (P=.34). The RSES fills an important gap in the existing toolkit of respiratory symptom questionnaires; it is a reliable and valid tool to assess respiratory symptoms in adult current and former smokers, including those who have switched to noncombusted nicotine products. This suggests that the scale is sensitive to respiratory symptoms that develop in smokers and to their remission when smokers quit or switch to noncombusted nicotine products intended to reduce the harm of smoking. The findings also suggest that switching from cigarettes to ENDS may improve respiratory health.

Item Response Theory Modeling of the Verb Naming Test.

Item response theory (IRT) is a modern psychometric framework with several advantageous properties as compared with classical test theory. IRT has been successfully used to model performance on anomia tests in individuals with aphasia; however, all efforts to date have focused on noun production accuracy. The purpose of this study is to evaluate whether the Verb Naming Test (VNT), a prominent test of action naming, can be successfully modeled under IRT and evaluate its reliability. We used responses on the VNT from 107 individuals with chronic aphasia from AphasiaBank. Unidimensionality and local independence, two assumptions prerequisite to IRT modeling, were evaluated using factor analysis and Yen's Q 3 statistic (Yen, 1984), respectively. The assumption of equal discrimination among test items was evaluated statistically via nested model comparisons and practically by using correlations of resulting IRT-derived scores. Finally, internal consistency, marginal and empirical reliability, and conditional reliability were evaluated. The VNT was found to be sufficiently unidimensional with the majority of item pairs demonstrating adequate local independence. An IRT model in which item discriminations are constrained to be equal demonstrated fit equivalent to a model in which unique discrimination parameters were estimated for each item. All forms of reliability were strong across the majority of IRT ability estimates. Modeling the VNT using IRT is feasible, yielding ability estimates that are both informative and reliable. Future efforts are needed to quantify the validity of the VNT under IRT and determine the extent to which it measures the same construct as other anomia tests. https://doi.org/10.23641/asha.22329235.

Investigating test content structure using multidimensional scaling

Although multidimensional scaling (MDS) has extensively been used in social and physical sciences to visualize and explore the latent dimensionality of observed data, few studies in language assessment have used MDS for the purpose of test construction, including construct and content validation. In this study, we use MDS to investigate and compare the similarity of writing prompts in the IELTS and TOEFL iBT tests. Random prompts from both tests were presented as item pairs and raters were asked to rate the similarity of the pairs in terms of content and cognitive complexity. The results showed the writing prompts in the TOEFL iBT test represented two major dimensions while those in the IELTS test demonstrated three domains.

Summarizing User-item Matrix By Group Utility Maximization

A user-item utility matrix represents the utility (or preference) associated with each (user, item) pair, such as citation counts, rating/vote on items or locations, and clicks on items. A high utility value indicates a strong association of the pair. In this work, we consider the problem of summarizing strong association for a large user-item matrix using a small summary size. Traditional techniques fail to distinguish user groups associated with different items (such as top- l item selection) or fail to focus on high utility (such as similarity- based subspace clustering and biclustering). We formulate a new problem, called Group Utility Maximization (GUM), to summarize the entire user population through k user groups and l items for each group; the goal is to maximize the total utility of selected items over all groups collectively. We show this problem is NP-hard even for l =1. We present two algorithms. One greedily finds the next group, called Greedy algorithm, and the other iteratively refines existing k groups, called k -max algorithm. Greedy algorithm provides the \((1-\frac{1}{e})\) approximation guarantee for a nonnegative utility matrix, whereas k -max algorithm is more efficient for large datasets. We evaluate these algorithms on real-life datasets.

Translation and psychometric properties of the Chinese version of the Sociocultural Attitudes Towards Appearance Questionnaire-4 (SATAQ-4) in college students

The current study aimed to translate the Sociocultural Attitudes Towards Appearance Questionnaire-4 (SATAQ-4) into Mandarin Chinese and to examine its psychometric properties. In Study 1 (N = 842, 427 women, 415 men), exploratory factor analysis supported a 4-factor solution for the Mandarin Chinese SATAQ-4, namely Internalization: Thin/Low body fat, Internalization: Muscular/Athletic, Pressures: Family/Peers, and Pressures: Media. Evidence also supported the high internal consistency and convergent validity of the male-subject and female-subject versions of Mandarin Chinese SATAQ-4. In Study 2 (N = 497, 290 women), confirmatory factor analyses supported the previously found 4-factor model, after allowing correlation between residuals of three pairs of items. In Study 3 (N = 294, 129 women), we examined the 3-month test-retest reliability across gender, and the results showed satisfactory ICC for both women (ICC =0.53–0.79) and men (ICC =0.56–0.66). Taken together, the male-subject and female-subject versions of the Mandarin Chinese SATAQ-4 are reliable and valid in evaluating sociocultural attitudes towards appearance.

Arcflow Formulations and Constraint Generation Frameworks for the Two Bar Charts Packing Problem

We consider the two bar charts packing problem (2-BCPP), a recent combinatorial optimization problem whose aim is to pack a set of one-dimensional items into the minimum number of bins. As opposed to the well-known bin packing problem, pairs of items are grouped to form bar charts, and a solution is only feasible if the first and second items of every bar chart are packed in consecutive bins. After providing a complete picture of the connections between the 2-BCPP and other relevant packing problems, we show how we can use these connections to derive valid lower and upper bounds for the problem. We then introduce two new integer linear programming (ILP) models to solve the 2-BCPP based on a nontrivial extension of the arcflow formulation. Even though both models involve an exponential number of constraints, we show that they can be solved within a constraint generation framework. We then empirically evaluate the performance of our bounds and exact approaches against an ILP model from the literature and demonstrate the effectiveness of our techniques on both benchmarks inspired by the literature and new classes of instances that are specifically designed to be hard to solve. The outcomes of our experiments are important for the packing community because they indicate that arcflow formulations can be used to solve targeted packing problems with precedence constraints and also that some of these formulations can be solved with constraint generation. History: Accepted by Andrea Lodi, Area Editor for Design &amp; Analysis of Algorithms–Discrete. Supplemental Material: The online supplement is available at https://doi.org/10.1287/ijoc.2022.1256 .

Development of an Inconsistent Responding Scale for the HEXACO Personality Inventory-Revised.

Inconsistent or careless responding is a significant threat to the validity of self-reported personality data. Using archival samples of undergraduate and community participants, we developed an inconsistent responding scale using items that appear on both the 60- and 100-item versions of the HEXACO Personality Inventory-Revised-two widely used measures of the HEXACO model of personality trait structure. We identified pairs of correlated HEXACO items in Sample 1 and created a total inconsistent responding score by summing absolute differences between each item pair. The Brief Response Inconsistency Evaluation (BRIE) for the HEXACO effectively differentiated between genuine and randomly generated responses across samples. The BRIE also correlated as expected with other measures of careless responding and relevant personality traits (e.g., conscientiousness). Tentative cut scores for the BRIE that appear to provide a reasonable balance between sensitivity and specificity in Sample 1 were investigated. Future research should examine the BRIE with different populations and translations of the HEXACO inventories and further investigate the effectiveness of the recommended cut scores.

Construction and preliminary validation of an inconsistent responding scale for the HEXACO-100.

Inconsistent responding poses a particular risk to self-report data and is a common phenomenon in undergraduate and community subject pools. The HEXACO-100 personality inventory (Lee & Ashton, 2018) lacks internal validity scales that could be used to identify inconsistent responding. To address this need, the present study developed a HEXACO Inconsistent Responding Test (HIRT). Highly correlated and conceptually overlapping HEXACO-100 item pairs were identified from a derivation sample and cross-validated by comparing real data to randomly generated protocols. The HIRT was highly accurate at identifying both randomly generated and half-randomly generated protocols. HIRT scores exhibited significant correlations with other inconsistency measures (i.e., the Triarchic Assessment Procedure for Inconsistent Responding and the HEXACO's Inter-item Standard Deviation) and were as effective as these other measures at differentiating between consistent and inconsistent responders. This preliminary validation of the HIRT provides strong evidence of its ability to improve the quality of data gathered with the HEXACO-100, which may be especially useful for researchers hoping to utilize the measure in clinically relevant research and applications. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Implicatures of proper name vocatives in English

Proper name vocatives sometimes induce implicatures. The phenomenon is probably well-known but not necessarily easy to explain in a pragmatically consistent fashion. The reason is that the same kind of name vocative yields implicatures in one occasion but sounds quite neutral in another. The goal of this paper is to advance a system by which we can predict when a given name vocative engenders implicatures in English. The system is based on the concept of conventional implicature, one of the three types of implicature delineated by Grice (1961, 1975, 1989). There is a special kind of conventional implicature inducer in which implicature is not directly encoded in a lexical item but in a contrastive pair of lexical items (like vous and tu in French). I propose that a set of competing proper name vocatives, namely, the full name, the (title plus) family name, the first name and the shortened first name (of a given addressee), is taken as one of those contrastive conventional implicature inducers. By capitalizing on Terkourafi's (2001; 2005a, 2008, 2009, 2012) concept of conventionalization (standardization), I develop a theory in which private level linguistic habits play a crucial role in inducing implicatures of proper name vocatives.