词汇识别中歧义词素语义加工:ERP研究

结合掩蔽启动范式和ERP技术,利用歧义词素的加工,本研究探讨了歧义词素语义的激活。实验采用3×2两因素被试内设计,歧义词素两含义互为启动刺激和目标刺激,两变量分别为启动词类型(歧义词素含义1 vs.含义2 vs.无关词)和目标词类型(歧义词素含义1 vs.含义2)。行为结果发现启动词类型和目标词类型存在交互作用,即只有当共享词素同形同义的情况下,目标词的加工才可能被启动,共享词素同形不同义时并没有产生启动效应。ERP结果进一步肯定了行为结果,在N250和N400两个脑电成分上都发现了启动类型和目标类型交互作用有关的效应,当启动词和目标词共享词素同形同义时,同时引发了N250和N400的变化;当共享词素同形不同义时,时程分析时虽然也发现了N250的变化,但是N250的效应较小,小于同形同义所引发的N250效应,持续时间也较短。实验结果说明,在词汇识别中歧义词素语义可以激活,并对目标词汇加工的早期加工阶段发挥作用。     

词素效应到底是什么？——来自维吾尔语的证据

利用维吾尔语的一些特点,采用视觉－视觉启动、听觉－视觉跨通道启动、和延迟重复启动等三种启动技术,改变目标词和启动词之间在词素、词形以及语义上的多种关系,通过比较启动效应的异同,来考察词素效应是否可以还原为词形启动效应、语义效应,是否可以看成是词形和语义在实时加工中相互作用的后果,从而探讨词汇加工中词素效应的实质。实验一发现,在明确配对的视觉－视觉和跨通道启动中,启动词和目标词之间只有在具有词素加语义关系（＋m＋o＋s / D,＋m＋o＋s / R;其中,m 代表词素,o 代表词形,s 代表语义,D 代表目标词为派生词,R代表目标词为自由词根;＋ 表示启动词和目标词之间在某个维度上相关,－表示启动词和目标词之间在某个维度上无关）时或者仅仅具有语义关系（－m－o＋s / D）时,才有显著的启动效应。如果启动词与目标词之间仅仅具有词素（词形）关系,而没有语义关系（＋m＋o－s / D）,或者仅仅具有词形关系（－m＋o－s / R）,它们之间的启动效应微乎其微。实验二通过使用维吾尔语中的借词和延迟启动技术,比较了词素加语义加词形（＋m＋o＋s / D）的启动和语义加词形启动（－m＋o＋s / D）以及语义启动（－m＋o＋s / D）,发现词素启动效应并没有不同于其他启动条件下的效应。两个实验都表明,语义关系是产生启动效应的关键,没有语义相关,就不可能产生词素启动,词素效应依赖于语义效应。有了语义关系,词素关系和词形关系产生的作用很小。因此,词汇加工的词素效应是词形和语义统计学规律的反应,是词形和语义加工相互作用的后果,心理词典中并不需要明确地表征词汇结构。     

二、三和五年级儿童对形旁语义的加工

结合启动范式和汉字判断任务,通过操纵目标字语义透明度,在汉字阅读过程探查了二、三、五年级儿童对形旁语义的加工。采用 3×2×2 混合实验设计,自变量为被试类型（小学二、三、五年级儿童）、启动字类型（字形相关 vs. 无关）和目标字类型（透明字 vs. 模糊字）。结果发现启动字类型、目标字类型和年级三者交互作用显著。随后分析在三五年级儿童中发现了显著的独立于字形的形旁语义启动效应,启动字中的共享形旁对模糊字的识别产生了更大的抑制作用,而在二年级儿童中并没有发现这种效应。说明形旁语义在儿童的心理词典中存在独立表征,可以被激活并促进儿童对汉字的识别,但是需要儿童的阅读理解能力发展到一定程度。     

重复启动作业中词的语义透明度的作用

用词素重复启动的方法考察合成词的心理表征。目标刺激为２４个语义透明度词、２４个不透明词以及４８个非词（都是两个字）。启动刺激有５种：（１）目标词本身,（２）目标词的第一词素,（３）目标词的第二词素,（４）由目标词的两个词素颠倒后形成的非词,（５）包含目标词的一个词素的非词。启动刺激的呈现时间是１００毫秒,随后是目标刺激４００毫秒。被试的任务是尽快准确地按键判断目标刺激是否是一个词。研究发现,（２）、（３）、（５）条件下的重复启动中,透明度表现出显著的效应：透明词的启动量显著大于不透明词。结果支持合成词的表征受语义透明度的影响,透明词和不透明词的词素和整词在心理表征中的关系可能不同：前者是一种兴奋性的连接,后者是一种抑制性的连接。     

语义透明度和构词频率对汉语动词多词素词识别的影响*

采用重复启动范式,通过两个实验,考察语义透明度和词素构词频率对动词多词素词识别的影响。结果发现：（1）语义透明度影响多词素动词的识别,识别高语义透明度的动词快于识别低语义透明度的动词。（2）词素的构词频率影响低语义透明度动词的识别,不影响高语义透明度动词的识别。对低语义透明动词,词素的构词频率低的动词的反应时显著短于词素的构词频率高的动词。研究结果支持汉语多词素词的多层次双系统选择加工模型。     

合成词加工中的词频、词素频率及语义透明度

通过变化汉语合成词的词素累积频率的方法,考察表面频率和语义透明度对加工速度的影响。研究发现,当控制透明度时,高表面频率词中存在词素累积频率效应,即词素频率越高反应越快;而在低表面频率词中没有。当控制表面频率,变化词素频率和透明度时,在透明词中,词素累积频率高的比低的加工快;但在不透明词中;这种反应趋势反转了：词素频率越高反应越慢。     

汉语逆序词的加工─—词素在词加工中的作用

四个实验以汉语道序同为实验材料,采用语义启动范式,考察了词素在汉语双字词识别中的作用及其在时间进程上的变化。实验一发现,在ＳＯＡ为１５７毫秒时,逆序词的呈现不但能促进与之有联想关系的目标词（语义启动）的词汇判断时间,而且也能促进其逆序形式的语义联想词（逆序启动）的词汇判断时间,两种启动效应之间没有差异。实验二和实验三则发现在ＳＯＡ分别为３１４毫秒和５７毫秒时,逆序词的呈现只对其联想词的词汇判断有显著的促进作用,而对其逆序形式的联想词的促进作用不显著。实验四在更为严格的实验条件下支持了实验三的结果。研究结论为：１）同素单元的激活在词的识别中起一定作用,但并不是词的识别的必经阶段;２）在时间进程上,词素单元的激活先是由弱到强,然后逐渐减弱,而整词单元的激活为稳定的上升过程;３）词素的位置信息的加工落后于词素单元本身的加工,发生在词汇加工的晚期。     

汉语字词识别中词频和语义透明度的交互作用

该研究通过词汇判断作业考察了中文双字合成词的心理表征问题。目标刺激为高频透明、高频不透明、低频透明、低频不透明词各12个和48个假词,被试任务为即既又准的判断目标刺激是否是一个词。结果发现,高频词的词汇判断速度快于低频词,透明词的词汇判断速度快于不透明词,词频和语义透明度有显著的交互作用。透明度对低频词的识别所起的作用大于对高频词的识别所起的作用。     

中文双字词的表征与加工(下)

2．3语义启动研究彭响龄，张必隐和刘志忠（1993）[1]采用语义启动的方法考察了中文双字词的词汇信息表征方式，所用的实验材料为联合式合成词。目标词与启动词之间在语义上有四种关系：1）整词相关，词素相关（如：杰出──优异）；2）整词相关，词素无关（如：恭维──奉承）；3）整词无关，词素相关（如：敲打──拍卖）；4）整词无关词素无关（如：耍弄──接替）。被试任务为通过按键进行词汇判断，对启动词和目标词都进行判断。从被试做出反应到下一个刺激呈现的间隔时间为240ms，具有语义关系的词对在实验材料中的比例为0.24。实验结…     

象似性和手语学习经验对中国手语-汉语跨模态语义启动效应的影响

采用启动条件下的词汇判断任务,考察熟练手语使用者和无手语经验成年听人的跨模态语义启动效应。结果发现：1)在象似词条件下,两组被试判断汉语语义相关词的反应时均快于语义无关词,说明手语象似词和汉语词之间存在跨模态语义启动效应。2)在非象似词条件下,仅手语熟练被试判断汉语语义相关词的反应时快于语义无关词,无手语经验被试判断汉语语义相关词和无关词的速度没有差异。这是由于前者心理词库中的手语词和口语词共享语义表征,而后者主要依赖手语象似词的视觉模拟性。整个研究表明,中国手语和汉语间存在跨模态语义启动效应,但该效应受到手语词象似性和手语学习经历的调节。     

Activation of morphemic meanings in processing opaque words

Two masked priming experiments were conducted to examine the activation of morphemic forms and meanings during opaque word processing. In Experiment 1, opaque primes significantly facilitated the recognition of transparent targets, which was consistent with previous results. However, transparent primes did not influence the recognition of opaque targets. This asymmetry could not be explained solely by morpho-orthographic processing, but it was consistent with models that have assumed early morpho-semantic activation. Experiment 2 directly tested whether the meanings of the constituent morphemes in opaque words were activated. In the critical condition, the targets were unrelated to the opaque primes at the lexical level, but were semantically related to their morphemes (e.g., “butterfly–bread”). Facilitation was observed in this condition, providing strong evidence of morpho-semantic activation during opaque-word recognition. These findings indicate that although initial morphological decomposition is determined by surface morphological form, it does not necessarily imply that morphemic meanings will be activated at later stages of processing. Rather, the morphemic meanings may be available automatically once segmentation is complete.     

Semantic aspects of morphological processing: transparency effects in Serbian

We examined the contribution of semantics to morphological facilitation in the visual lexical decision task at two stimulus onset asynchronies (SOAs) with Serbian materials. Primes appeared in Roman or Cyrillic characters. Targets always were printed in Roman. When primes were presented at an SOA of 250 msec, decision latencies to verbal targets (e.g., VOLIM) showed greatest facilitation after inflectionally (e.g., VOLE) related primes, significantly less after semantically transparent derived primes (e.g., ZAVOLE), and less again after semantically opaque derived primes (e.g., PREVOLE). Latencies after semantically transparent and opaque derived target words did not differ at an SOA of 48 msec. Both were slower than after inflectionally related primes. Stated generally, effects of semantic transparency among derivationally related verb forms were evident at long SOAs, but not at short ones. Under alphabet-alternating conditions, magnitudes of facilitation were greater overall, but the pattern was similar. The outcome suggests that restricted processing time for the prime limits the contribution of semantics to morphological processing and calls into question accounts that posit a task-invariant semantic criterion for morphological decomposition within the lexicon.     

Semantic transparency and masked morphological priming: The case of prefixed words

In four lexical decision experiments, we investigated masked morphological priming with Dutch prefixed words. Reliable effects of morphological relatedness were obtained with visual primes and visual targets in the absence of effects due to pure form overlap. In certain conditions, priming effects were significantly greater with semantically transparent prefixed primes (e.g., rename-name) relative to the priming effects obtained with semantically opaque prefixed words (e.g., relate-late), even with very brief (40-msec) prime durations. With visual primes and auditory targets (cross-modal priming), significant facilitation was found in all related prime conditions, independent of whether or not primes were morphologically related to targets. The results are interpreted within a bimodal hierarchical model of word recognition in which morphological effects arise through the interplay of sublexical (morpho-orthographic) and supralexical (morpho-semantic) representations. The word stimuli from this study may be downloaded as supplemental materials from http://mc.psychonomic-journals .org/content/supplemental.     

What do graded effects of semantic transparency reveal about morphological processing?

We examined the influence of semantic transparency on morphological facilitation in English in three lexical decision experiments. Decision latencies to visual targets (e.g., CASUALNESS) were faster after semantically transparent (e.g., CASUALLY) than semantically opaque (e.g., CASUALTY) primes whether primes were auditory and presented immediately before onset of the target (Experiment 1a) or visual with an stimulus onset asynchrony (SOA) of 250 ms (Experiment 1b). Latencies did not differ at an SOA of 48 ms (Experiment 2) or with a forward mask at an SOA of 83 ms (Experiment 3). Generally, effects of semantic transparency among morphological relatives were evident at long but not at short SOAs with visual targets, regardless of prime modality. Moreover, the difference in facilitation after opaque and transparent primes was graded and increased with family size of the base morpheme.     

Transposed-Letter Priming Across Inflectional Morpheme Boundaries

This study reports findings from two experiments testing whether a transposed-letter (TL) priming effect can be obtained when the transposition occurs across morphological boundaries. Previous studies have primarily tested derivationally complex words or compound words, but have not examined a more rule-based and productive morphological structure, i.e., inflectionally complex words, using masked priming. Experiment 1 tested TL priming with nonword primes and inflected targets (FOCUSING). Nonword primes were formed by transposing letters either within the root morpheme (fcousing) or across two morphemes (focuisng). Experiment 2 used the same nonword primes, but had the root words as targets (FOCUS). Both experiments showed similar TL priming effects for within-morpheme and across-boundary positions, indicating that morphological decomposition takes place only after letter positions in a word have been assigned. This finding provides additional evidence to previous research testing derived and compound words showing TL priming regardless of the position of transposition.     

Early morphological processing is morphosemantic and not simply morpho-orthographic: A violation of form-then-meaning accounts of word recognition

Many studies have suggested that a word’s orthographic form must be processed before its meaning becomes available. Some interpret the (null) finding of equal facilitation after semantically transparent and opaque morphologically related primes in early stages of morphological processing as consistent with this view. Recent literature suggests that morphological facilitation tends to be greater after transparent than after opaque primes, however. To determine whether the degree of semantic transparency influences parsing into a stem and a suffix (morphological decomposition) in the forward masked priming variant of the lexical decision paradigm, we compared patterns of facilitation between semantically transparent (e.g., coolant—cool) and opaque (e.g., rampant—ramp) prime—target pairs. Form properties of the stem (frequency, neighborhood size, and prime—target letter overlap), as well as related—unrelated and transparent—opaque affixes, were matched. Morphological facilitation was significantly greater for semantically transparent pairs than for opaque pairs. Ratings of prime—target relatedness predicted the magnitude of facilitation. The results limit the scope of form-then-meaning models of word recognition and demonstrate that semantic similarity can influence even early stages of morphological processing. The research reported here was supported by National Institute of Child Health and Development Grant HD-01994 to Haskins Laboratories.     

Morphological processing during visual word recognition in developing readers: evidence from masked priming

Masked priming studies with adult readers have provided evidence for a form-based morpho-orthographic segmentation mechanism that "blindly" decomposes any word with the appearance of morphological complexity. The present studies investigated whether evidence for structural morphological decomposition can be obtained with developing readers. We used a masked primed lexical decision design first adopted by Rastle, Davis, and New (2004), comparing truly suffixed (golden-GOLD) and pseudosuffixed (mother-MOTH) prime-target pairs with nonsuffixed controls (spinach-SPIN). Experiment 1 tested adult readers, showing that priming from both pseudo- and truly suffixed primes could be obtained using our own set of high-frequency word materials. Experiment 2 assessed a group of Year 3 and Year 5 children, but priming only occurred when prime and target shared a true morphological relationship, and not when the relationship was pseudomorphological. This pattern of results indicates that morpho-orthographic decomposition mechanisms do not become automatized until a relatively late stage in reading development.     

Morphological analysis by child readers as revealed by the fragment completion task

Ten-year-old children performed a fragment completion task. Target fragments (e.g., T_ _N) were preceded by four types of study conditions. The identity condition consisted of the target (TURN). Themorphological condition included a related form (TURNED). The orthographic condition consisted of morphologically unrelated words (e.g., TURNIP). Finally, no similar word was presented in the study phase of the no-prime condition. Morphological relatives included orthographically transparent (TURNED-TURN) and orthographically opaque (RIDDEN-RIDE) forms. The results indicated that performance of child readers on the fragment completion task was sensitive to morphological relationships. Completion rates following opaque, as well as transparent, morphological relatives were significantly greater than those following orthographically similar forms. In sum, the fragment completion task provides a viable new tool for examining morphological processing in children and for differentiating morphological effects from effects of similar form.